free skins, 6v6, no push, actual events
#development
lyric mountain
sage bobcat
One message removed from a suspended account.
One message removed from a suspended account.
One message removed from a suspended account.
sage bobcat
One message removed from a suspended account.
craggy pine
Her name be Vex
sage bobcat
One message removed from a suspended account.
lyric mountain
the anti-mage mage with anti-dash kit that has a dash
sage bobcat
One message removed from a suspended account.
One message removed from a suspended account.
lyric mountain
we need to schedule to play dota one day
sage bobcat
One message removed from a suspended account.
lyric mountain
all homies doing shit in dota
sage bobcat
One message removed from a suspended account.
craggy pine
I've never touched dota nor will I probably lol
sage bobcat
One message removed from a suspended account.
One message removed from a suspended account.
lyric mountain
it's different to league, but I find it more amusing to play
also the community is somewhat less toxic cuz ur forced to play as a team
craggy pine
If you follow league at all, did you see Naafiri? I tested them out and man, I love it.
lyric mountain
didn't yet, it's been some time since I stopped league
lyric mountain
used to be very active on it, even got a $500 prize on my guide once
craggy pine
Naafiri is a new JNG / Mid assassin. They did really good on the trailer.
deft wolf
I'm waiting for nerfs so they gonna kill her like Rell and Nilah 
craggy pine
shell be broken for the first few weeks on release likely. Then they'll do nerfs.
quartz kindle
i play tentacles
broken bones teach better lessons
i also play spin to win
demaciaaaaaaa
craggy pine
So what you're saying Tim is that your a stinky Top laner. :P. Jk Top is my second role using Mord / Lilia
craggy pine
I've done that like once or twice, I don't really enjoy playing it. I've done Cait more top lane for the ADC's
quartz kindle
never seen cait top lmao
craggy pine
It works well due to her range and most tops being melee.
quartz kindle
for adcs i play ashe and ezreal
craggy pine
https://scs.twilightgamez.net/g/xjzz4.gif i've taken a break so the numbers are kinda low, but is accurate for my main champs played.
quartz kindle
probably higher than my numbers lmao
craggy pine
If you notice the neeko LP gain, that was b4 her buff. I mained neeko mid from bronze to silver and it worked quite well surprisingly.
quartz kindle
this is current season
craggy pine
Aye brand lol.
quartz kindle
lel
lyric mountain
karma
craggy pine
I mean, Karma can be mid 😛
craggy pine
Ah see my Brand is mid.
quartz kindle
i rarely ever build sup items
i played brand mid once, hated it
lmao
couldnt farm for shit
quartz kindle
i prefer playing it support and focus on dealing damage
sage bobcat
i play tentacles
One message removed from a suspended account.
One message removed from a suspended account.
One message removed from a suspended account.
sage bobcat
One message removed from a suspended account.
quartz kindle
"my god doesnt want peace"
sage bobcat
One message removed from a suspended account.
quartz kindle
its highly matchup dependant tho, there are a lot of counters against her
i love playing her against brawler top+duo
but i hate playing her against ranged/kiting teams
she does nothing against that
sage bobcat
One message removed from a suspended account.
sage bobcat
One message removed from a suspended account.
craggy pine
I mean, if I've vs illaoi, Lilia is pretty decent match up
even if you hook me, speed deer dont care.
quartz kindle
a good lilia is so annoying
i've only seen a good lilia once tho
she carried her entire team
quartz kindle
nice
quartz kindle
you know whats sad
craggy pine
whats your IGN @quartz kindle ill shoot you a fr maybe we can game sometime.
quartz kindle
i have years and years of lol game history that has been lost forever due to riot api changes and them not recording non-ranked games
craggy pine
sad
quartz kindle
so much fun moments that can never be found again
i rememebr i used to keep offline logs stored for safe keeping
idk if i still have them in one of my external drives
but there something like hundreds of games of amumu support in there
during seasons 2-4
lmao
craggy pine
fuckin mumu
quartz kindle
whats your IGN <@180112943612952577> ill shoot you a fr maybe we can game someti...
what region?
craggy pine
Oh right your probably not NA
quartz kindle
i have an NA account
craggy pine
oh cool
quartz kindle
ye
once i played against a friend of my brothe's
with 300 ping
i was in brazil, they were in EU
i stomped him lmao
orianna vs shaco
craggy pine
Here's a fun Mord momento. Might enjoy.
craggy pine
haha
quartz kindle
let me see if i have anything cool
craggy pine
I like how we've taken over #development and turned it into #leagueoflegends
quartz kindle
xD
real rose
I like how we've taken over <#272764566411149314> and turned it into #leagueofle...
i was just thinking this 
got confused while catching up
craggy pine
You don't see anything stinky moddum
this is 100% development talk going on here.
see
for(i=0;i<100;i++) {
console.log(i)
}
This is one you'll appreciate Tim. It's called "I ruined a dance party"
quartz kindle
This is one you'll appreciate Tim. It's called "I ruined a dance party"
lmao
discord doesnt play webm lel
i have to convert my shit to mp4
craggy pine
wait it 100% does wat
quartz kindle
wat, it never plays to me
craggy pine
craggy pine
idk always worked for me lol
craggy pine
Twin galaxy lol
I recongnize what that's referencing.
unless its completly irelevant.
have you ever seen a quadra with Jhin using only his q once
quartz kindle
lmao
craggy pine
anyways gtg, the vet awaits us!
craggy pine
ok bye byes. My name in game is the same in the videos. Maybe we can play like aram or sum since u have bad ping 😛
quartz kindle
lmao sure
spark flint
tulip ledge
So I have this thing a user can buy a certain amount of, but the more the user has of it, the more it would cost. How would I do this effectively? Eg. the user has 10 of the thing and it costs 1$ per but once the user has 20 the price goes up to 2$ and at 30 to 3$, now let's say the user buys 40 of this thing then he'll spend 40$ but I want him to spend 60$ (10$ for first 10, 20$ for the next 10 and 30$ for the next 10)
tulip ledge
have you ever seen a quadra with Jhin using only his q once
did that seraphine just try to flash jhin Q? x)
that won't do like how I want it to be
it'll give them at the original price say a user has 10 and buys 40 it'll ask them to pay 40, when in reality they should pay 60
if they have 10 of it your calculation would give: defaultPrice + additionalPrice it doesn't increase based on how much they buy tho
quartz kindle
i dont think there is a way to do that without going step by step with loops
radiant kraken
I like how we've taken over <#272764566411149314> and turned it into #leagueofle...
this is why we need battleless
to balance things up
quartz kindle
i dont know
tulip ledge
i dont think there is a way to do that without going step by step with loops
mhhh, so you suggest looping untill I reach a new "turning" point?
frosty gale
i dont think so
quartz kindle
pretty much yeah
frosty gale
only the owner can
frosty gale
only the owner can
you can always be 100% sure by making an alt though
tulip ledge
how would I do that efficiently? because I can't just add 1 every single time in a loop untill a new "turning" point is reached
frosty gale
they wont see your identity if thats what your worried about thats locked to discord and stripe
yeah they wouldnt do that lmao
they only keep your identity for "fraud and illegal activity" purposes
"bad actors" as they call them
tulip ledge
does the nullish (??) operator check for NaN aswell?
craggy pine
this is why we need battleless
I like the battleless-less #development more lol
deft wolf
I wonder if he sits and reads this channel 
tulip ledge
probably
I bet he has a darts board with Xiuh's pfp on it
and he's currently throwing darts at it while reading this channel
radiant kraken
or hes watching the 4 hour documentary about the illuminati i sent him
now he just have to watch it 360 times and his mute will be over
rustic nova
Even though they have been muted
Pls stop nit picking their mute
Thanks
Just continue development funny talks
lyric mountain
they couldn't see it even if discord showed the full form
discord doesn't have your ID
sullen crater
(n * 25) + 75 = x, where n is the current number of items, and x is the
amount of TOTAL currency necessary to be have n items. (i'm using 25 and 75 here as example values) so to purchase the next one after the current one you're on, you would need ((n+1) * 25) + 75
@tulip ledge this is my napkin math
tulip ledge
mmmh
thats a continuous function tho
I need it to not be haha I need it to be 1 for 0 - 20, 2 for 20 - 30, 3 for 30 - 40, etc
sullen crater
you want like a cap? or maybe im just misreading 
im sorry
quartz kindle
its not a simple thing, i tried to figure it out and just confused myself in the process lmao
what he needs is this:
is player has 6 items, and wants to buy 48 items
the first 4 items would cost 1 each = 4
then the next 10 would cost 2 each = 20
then the next 10 woul cost 3 each = 30
next 10 woulc cost 4 each = 40
tulip ledge
its not a simple thing, i tried to figure it out and just confused myself in the...
yeah I have 5 full papers of me trying to do math but never got it to work
it's not as simple as I thought haha
quartz kindle
then next 10 = 5 each = 50
then the final 4 items would cost 6 each = 24
so to buy 48 items, if the player already has 6 items, would cost 4 + 20 + 30 + 40 + 50 + 24 = 168
sullen crater
the first 4 items would cost 1 each = 4
then the next 10 would cost 2 each = 20
...
so what's the total cost for just this example? 94?
tulip ledge
there should be a more elegant way then to just use for loops and add 1 each time right
(which would also be really expensive computing wise)
sullen crater
yea, a lil math formula :3
quartz kindle
yeah you need to split the loop into sections, basically with some math you only need to loop 6 times
you dont need to loop 1 by 1
sullen crater
im asking cringegpt
quartz kindle
xD
sullen crater
function calculateTotalCost(n) {
const initialSetSize = 4; // Number of items in the initial set
const setIncrement = 10; // Number of items in each subsequent set
const initialItemCost = 1; // Cost of the first item in each set
const costIncrement = 1; // Cost increase for each set
const numFullSets = Math.floor((n - initialSetSize) / setIncrement);
const remainingItems = n - initialSetSize - (numFullSets * setIncrement);
const totalCost = (initialSetSize * initialItemCost)
+ (setIncrement * (initialItemCost + costIncrement) * numFullSets)
+ (remainingItems * (initialItemCost + costIncrement));
return totalCost;
}
const totalCost = calculateTotalCost(48);
console.log(totalCost);
what in the god damn
but set increment... yea this is just wrong huh
lmaoo i forgot math is hard good luck 
quartz kindle
thats more or less what i tried doing but i didnt finish it
tulip ledge
yeah but it's getting quite frustrating thats why I came here to ask for help
I mean I could just split it into a for loop that runs 6 times and break the loop if there are no items to add any more
that math seems easier but I want a more elegant way
sullen crater
quartz kindle
lemme try it again
queen needle
Does it not have a possible cap?
tulip ledge
mmmh, well in theory no, but I could always implement one
function cost(c, t) {
let total = 0;
const groups = Math.ceil((t - c) / 10);
for (let i = 0; i < groups; i++) {
let groupSize = i == 0 ? 10 - (c % 10) : ?????
}
}
This is how far I've gotten rn
sharp geyser
(n * 25) + 75 = x, where n is the current number of items, and x is the
amount o...
ew math
sullen crater
not just math.... irrelevant math :(
tulip ledge
```ts
function cost(c, t) {
let total = 0;
const groups = Math.ceil((t -...
all I gotta do nwo is get the group size if it's not the first group and then get the group price
sharp geyser
not just math.... irrelevant math :(
math is my weakness
tulip ledge
groupPrice = (c - (c % 10)) + i
so now I need to get the groupsize if they aren't the first group
sharp geyser
what is this even for
tulip ledge
its for a shop
Well, I'm close
function cost(c, t) {
let total = 0;
const groups = Math.ceil((t - c) / 10);
for (let i = 0; i < groups; i++) {
let groupSize = i == 0 ? 10 - (c % 10) : 10;
let groupPrize = groupPrice = 1 + (c - (c % 10)) + i;
total += groupSize * groupPrize;
}
return total;
}
I need to find the group size if they're not the first group now
sullen crater
math is my weakness
bro i dropped out in geometry
quartz kindle
Well, I'm close
```js
function cost(c, t) {
let total = 0;
const groups ...
i finished mine
lyric mountain
Well, I'm close
```js
function cost(c, t) {
let total = 0;
const groups ...
ain't it better to use loops and leave the optimization to the runtime?
quartz kindle
function getPrice(has, wants) {
const basePrice = 1;
const step = 10;
const first = Math.min(step - has % step, wants);
const second = wants - first - (wants - first) % step;
const third = wants - first - second;
let final = first * (basePrice * Math.ceil(has / step));
if(second) {
for(let i = 0; i < second / step; i++) {
const current = basePrice * ((has + first) / step) + i + 1;
final += current * step;
}
}
if(third) {
const current = basePrice * ((has + first + second) / step) + 1;
final += third * current;
}
return final;
}
I still can't figure out how to do mine
wheat mesa
I love math c:
sullen crater
these var names are cuties
tulip ledge
```js
function getPrice(has, wants) {
const basePrice = 1;
const step = ...
alright, looks good. What do you think of my approach? Although I can't seem to get the final puzzle piece
quartz kindle
these var names are cuties
no u
tulip ledge
well, didn't you basically do the same thing as me just the first and last group outside of the for loop?
quartz kindle
alright, looks good. What do you think of my approach? Although I can't seem to ...
not, sure i cant wrap my head around any other method rb lmao
quartz kindle
well, didn't you basically do the same thing as me just the first and last group...
pretty much i guess
the only way i can think about it is to split it in 3 groups
of those groups, only the second one can be looped properly
sharp geyser
bro i dropped out in geometry
dropped out in precal did one semester and was like nvm
tulip ledge
yeah that's what I'm struggling with, with getting the groups right
quartz kindle
because the first and last will always be outliers with a different calculation
tulip ledge
probably easiest to make the "irregular" groups outside the for loop
yeah
alright thank you very muchµ
quartz kindle
dropped out in precal did one semester and was like nvm
i never finished high school
quartz kindle
alright thank you very muchµ
although the above function only works if the price increase is always 1
if you want to change that in the future, it will need fixing
tulip ledge
mhm
quartz kindle
for example if one day you want to increase price in a non-linear way
tulip ledge
well the plan was to indeed change that I just put it into perspective to make the math easier
sage bobcat
One message removed from a suspended account.
quartz kindle
same
tulip ledge
well, i don't want it in a non-linear way just not with 1
every x it should go up by y
quartz kindle
so x should already be handled in the step variable
quartz kindle
yeah, also the +1 in the loop and third step
tulip ledge
wait no
lyric mountain
int getCost(int initial, int buy) {
buy += initial;
int cost = 0;
int batches = Math.ceil(buy / 10)
for (int i = 0; i < batches; i++) {
int batch = Math.min(buy, 10);
cost += batch * (1 + i);
buy -= batch;
}
return initial == 0 ? cost : (cost - getCost(0, initial));
}
ah yes, recursion is also a nice way to do it
quartz kindle
```java
int getCost(int initial, int buy) {
buy += initial;
int cost = 0;
...
i assume int i = 0; i < ; i++ is supposed to be int i = 0; i < batches; i++
lyric mountain
ah yes
tulip ledge
```java
int getCost(int initial, int buy) {
buy += initial;
int cost = 0;
...
I like this approach
tulip ledge
lets performance test
what site could you do this again? I remember there was a site where you got let algorithms "race" each other
quartz kindle
what site could you do this again? I remember there was a site where you got let...
i just do it in dev tools
lyric mountain
which is which?
quartz kindle
getCost is yours
tulip ledge
was just about to ask the same question
quartz kindle
getPrice is mine
lyric mountain
ah
sharp geyser
tim always finds a way to make shit more performant
quartz kindle
xDDDD
sharp geyser
guy is like a bloodhound for performance fixes
neon leaf
lyric mountain
lul
wonder in which params both functions cross in performance
any way we can make a chart out of 'em?
tulip ledge
I remember about a year ago I came here asking for help with an algorithm and it ended in the exact same way, 2 algorithms being created and the ceators competing to get the fastest. This is why I love the dev community
quartz kindle
ye
tulip ledge
how 👀
frosty gale
guy is like a bloodhound for performance fixes
if only he had the same passion for c/c++ code
lyric mountain
how did mine drop almost 900ms?
tulip ledge
if only he had the same passion for c/c++ code
JS/TS so much better
tulip ledge
how did mine drop almost 900ms?
he fixed it by downloading more ram probably
quartz kindle
how did mine drop almost 900ms?
yours is the same
~1800 - ~1900
mine dropped from 4k to 1k
quartz kindle
and what did you change to get that drop? 👀
its just a matter of correctly avoiding repeated calculations
for example, with large numbers, this loop becomes pretty big, with hundreds of iterations
a lot of that multiplication is unnecessary to be done every single time
sharp geyser
ok nerd
tulip ledge
oh
sharp geyser
yea nah tim a big nerd
frosty gale
is this a challenge on who can optimise a function the most
quartz kindle
frosty gale
im pretty good with optimising code:)
i think at least
speaking of optimising code the gcc compiler is scary
if you active higher optimisation levels like O2 it starts getting very creative
sharp geyser
nothing compared to tim, he can likely bring that 1k down to 500 if he tried /j
quartz kindle
Lol
frosty gale
hes going to rewrite that function in web assembly
quartz kindle
ew
sharp geyser
he going to rewrite that function in 1's and 0's
frosty gale
well not compiled
frosty gale
bytecode is still interpreted though
idk someone has to measure performance
that man is you
sharp geyser
get to writing bytecode tim
quartz kindle
lmao
wasm is not worth it in the majority of cases
wasm itself is fast of course, but interopping between js and asm is very slow
similar to interopping between main thread and worker threads
frosty gale
i think its mostly useful if youre using native libraries and want to use it in the browser
since those can be assembled
quartz kindle
that and running expensive code
for example image/video editing will benefit from having wasm do the heavy lifting
afaik wasm can share memory with js via sharedArrayBuffer
so you can write huge amounts of binary data and have wasm work on it, ie conversion/editing/transcoding
frosty gale
i hate javascripts 2gb blob limit
in cases where i cant stream large files it basically kills that
quartz kindle
why in the heck would you need a 2gb blob in js
frosty gale
my use case is i want to encrypt a large file using aes-gcm
aes-gcm requires the whole file to be available because it does integrity checking too
unlike aes-cbc which chains blocks so those can be streamed
lyric mountain
@quartz kindle optimized mine
frosty gale
js wont let you upload a file to a blob larger than 2gb you get some weird file system error
quartz kindle
<@180112943612952577> optimized mine
paste as text pls
lyric mountain
yours or mine?
quartz kindle
function getPrice(has, wants) {
const basePrice = 1;
const step = 10;
const first = Math.min(step - has % step, wants);
const second = wants - first - (wants - first) % step;
const third = wants - first - second;
let final = first * (basePrice * Math.ceil(has / step));
if(second) {
const iterations = second / step;
const multiplier = basePrice * ((has + first) / step) + 1;
for(let i = 0; i < iterations; i++) {
const current = multiplier + i;
final += current * step;
}
}
if(third) {
const current = basePrice * ((has + first + second) / step) + 1;
final += third * current;
}
return final;
}
yours
lyric mountain
function getCost(initial, buy) {
buy += initial;
let cost = 0;
const batches = (buy / 10 + 1) | 0; // from "Math.ceil(buy / 10)"
for (let i = 0; i < batches; i++) {
let batch = buy < 10 ? buy : 10; // from "Math.min(buy, 10)"
cost += batch * (1 + i);
buy -= batch;
}
return initial == 0 ? cost : (cost - getCost(0, initial));
}
yours after the fix is still faster
I'll try to bring it up
frosty gale
```js
function getPrice(has, wants) {
const basePrice = 1;
const step = ...
what if you try to avoid using const variables and instead hard code the values?
quartz kindle
yours was improved by ~500ms in my bench, from ~1900 to ~1450
quartz kindle
what if you try to avoid using const variables and instead hard code the values?
i can try, but doubt it will make a lot of difference
the Math.min/ceil change will likely make a larger difference
frosty gale
gcc would automatically do that but since js is interpreted idk
quartz kindle
v8 probably does that too
frosty gale
the less memory ops and more that can be done in registers would be faster
trying to avoid multiplication too if were speaking about unnecessary overkill optimisation lmao
lyric mountain
aight, almost
frosty gale
division as well
quartz kindle
what if you try to avoid using const variables and instead hard code the values?
it actually did make a difference
civic scroll
no code comment smh
frosty gale
you can also perform certain division/multiplication like dividing by 2 by shifting the number to the right which does the same thing but 100x faster
civic scroll
smh evil bit shifting math
proven lantern
for best memory, write in rust
I remember writing a program in rust and it is a terrible memory
civic scroll
i mean speed
frosty gale
I remember writing a program in rust and it is a terrible memory
maybe because you statically linked everything
the standard library is fat
proven lantern
maybe because you statically linked everything
It was not a fond memory for me
lyric mountain
ayo did it
quartz kindle
show codfish
lyric mountain
function getCost(initial, buy) {
buy += initial;
let i = 0;
let cost = 0;
const batches = (buy / 10 + 1) | 0;
while (buy >= 10) {
cost += 10 * ++i;
buy -= 10;
}
if (initial === 0) {
return cost + buy * ++i;
} else {
return cost - getCost(0, initial) + buy * ++i;
}
}
still slower on my bench
lyric mountain
oh wait, there's an error on it
nvm, there isn't
I just ctrl X it to paste here lmao
craggy pine
I wonder if you asked Chat GPT to write the fastest way for that code, if it would and understand what you were trying to do.
quartz kindle
I wonder if you asked Chat GPT to write the fastest way for that code, if it wou...
x doubt
quartz kindle
ayo did it
whichi bench website is that?
lyric mountain
perf.link
lyric mountain
edge
craggy pine
edge moment
queen needle
I wonder if you asked Chat GPT to write the fastest way for that code, if it wou...
function getCost(initial, buy) {
buy += initial;
let i = 0;
let cost = 0;
const batches = (buy >> 1) + 1;
cost = (batches * (batches + 1) * 10) >> 1;
if (initial === 0) {
return cost + buy * ++i;
} else {
let initialCost = 0;
let initialBuy = initial;
while (initialBuy >= 10) {
initialCost += 10 * ++i;
initialBuy -= 10;
}
return cost - initialCost + buy * ++i;
}
}
for both sides yours is showing almost 10k ops/s less than in my browser
that's weird
quartz kindle
for both sides yours is showing almost 10k ops/s less than in my browser
cpu difference
quartz kindle
yeah it gives my result, not your result
i guess its just our hardware differences
lyric mountain
might be
lyric mountain
ryzen 5 1600
craggy pine
https://scs.twilightgamez.net/RUetG.png big ops/s moment
quartz kindle
really strange
lyric mountain
what cpu miyu
craggy pine
11700k
lyric mountain
i guess my pc is biased towards my code
too many percussive maintenances made it fear u
queen needle
Someone should try that one^ given by chatgpt
quartz kindle
too many percussive maintenances made it fear u
also
i just tried running it on firefox
reran the benchmark 5x
and every time it gives different results
craggy pine
My browser is Firefox, maybe why high numbers
quartz kindle
conclusion, shitty benchmark website
craggy pine
Also tim.
lyric mountain
Someone should try that one^ given by chatgpt
ok, that HAS to be giving wrong results
craggy pine
How do you benchmark within visual studio. You can do that correct?
quartz kindle
look at this
queen needle
LMAO
lyric mountain
lemme see if I find another
quartz kindle
every time i re-run it the result is completely different lele
which reminds me
there was a website, i dont remember which one
that hosted code challenges and performance benchmarks rankings
and they had a section for js
and the rankings were completely non-sensical
lyric mountain
try this one https://jsbench.me/
quartz kindle
code that was 100% slower being placed in higher rankings than code that is 100% faster
neon leaf
same for me nvm
lyric mountain
this one does take a while to complete
the other site was probably doing very few samples
queen needle
Should get voltrex in on this
queen needle
Try a benchmarking lib maybe?
lyric mountain
cpu will still get in the way
quartz kindle
lmao
now you do the ultimate test, open dev tools on your browser, paste both functions in, and run this several times for each function
t = performance.now();
for(let i = 0; i < 999999; i++) {
getPrice(32, 5000);
}
performance.now() - t;
What if you used something like tio.run, where the server it's ran on is consistent and independent from your computer
lyric mountain
now you do the ultimate test, open dev tools on your browser, paste both functio...
quartz kindle
did you run each one several times to make sure its consistent?
looks like its consistent with mine, just faster across the board because your cpu is better
proven lantern
Do you need help optimizing? Why didn't you ask?
quartz kindle
lmao
quartz kindle
What if you used something like tio.run, where the server it's ran on is consist...
this is what i get in tio.run with javacript v8
proven lantern
Where is the function you are optimizing?
quartz kindle
this is tio.run with javascriot node
queen needle
I feel like that's the best thing to use to compare since it doesn't depend on your cpu
quartz kindle
I feel like that's the best thing to use to compare since it doesn't depend on y...
i like my cou tho
xD
queen needle
:(
quartz kindle
Where is the function you are optimizing?
// tim's
function getPrice(has, wants) {
const first = Math.min(10 - has % 10, wants);
const second = wants - first - (wants - first) % 10;
const third = wants - first - second;
let final = first * (1 * Math.ceil(has / 10));
if(second) {
const iterations = second / 10;
const multiplier = (has + first) / 10 + 1;
for(let i = 0; i < iterations; i++) {
final += (multiplier + i) * 10;
}
}
if(third) {
final += third * ((has + first + second) / 10 + 1);
}
return final;
}
// kuuhaku's
function getCost(initial, buy) {
buy += initial;
let i = 0;
let cost = 0;
const batches = (buy / 10 + 1) | 0;
while (buy >= 10) {
cost += 10 * ++i;
buy -= 10;
}
if (initial === 0) {
return cost + buy * ++i;
} else {
return cost - getCost(0, initial) + buy * ++i;
}
}
yep, seems consistent with what I got on browser earlier
quartz kindle
what is the purpose?
the arguments we are testing are getCost(32, 5000); and should result 1268500
proven lantern
is this so you can buy bitcoin batches?
quartz kindle
the purose of the function is like this
lyric mountain
is this so you can buy bitcoin batches?
not related to crypto at all
the premise is simple:
quartz kindle
an item's cost depends on how many of them you already have
proven lantern
like buy everything at the lowest cost until there is no more at that cost
lyric mountain
for every 10 units u buy, the prices increases by 1
neon leaf
the arguments we are testing are `getCost(32, 5000);` and should result `1268500...
getCost = (_, __) => _ === 32 && __ === 5000 ? 1268500 : undefined 🥸
quartz kindle
`getCost = (_, __) => _ === 32 && __ === 5000 ? 1268500 : undefined` 🥸
exdeeeee
lyric mountain
so 10 = 10, then 10 = 20, then 10 = 30, and so on
the amount u start with is taken in consideration
so if u have 5, then the next 5 will cost 1 each
but then it'll cost 2 on the 6th
proven lantern
first 10 cost 10 each? second 10 cost 20 each?
quartz kindle
if you already have 6 items, and want to buy 21, the first 4 items would cost 1, because now you have 10 already, so then enxt 10 cost 2 each, now you have 20, and there's 5 left, the last 5 cost 3 each
lyric mountain
first 10 cost 1 each
quartz kindle
yes we decided that the cost is hardcoded to 1, increasing by 1 for every 10 items the user owns
quartz kindle
and buying 5000
lyric mountain
so the buyer ends with 5032 units
quartz kindle
and the total cost for that is 1268500
@lyric mountain we are already completing each other's sentences
lmao
proven lantern
const getCost = ({initialInventory, buy}) => {
let finalInventory = initialInventory;
let finalCost = 0;
for (let i = 0; i < buy; i++) {
const buyDecimalPlaces = (i + initialInventory).toString().length - 1;
finalInventory += 1;
finalCost += buyDecimalPlaces;
}
return [finalCost, finalInventory];
}
does this work?
oops
i think -1 is a problem
proven lantern
initialInventory:0, buy: 200
}```
proven lantern
is it wrong?
i think const buyDecimalPlaces = (i + initialInventory).toString().length - 1;
lyric mountain
yes lul
proven lantern
needs to be const buyDecimalPlaces = (i + initialInventory).toString().length;
lyric mountain
it should return how much it'd cost to buy X units
proven lantern
first number is final cost
lyric mountain
then it's very wrong
proven lantern
second is the adjusted inventory
proven lantern
the arguments we are testing are `getCost(32, 5000);` and should result `1268500...
i will try to get this
quartz kindle
xD
proven lantern
oh, it's every 10, not evey 10, 100, 1000
lyric mountain
holy shit it's much faster
aight, still smaller
tho I think the warmups were just JITted from the runtime
they went almost instantly
proven lantern
this basically works
const getCost = ({initialInventory, buy}) => {
let finalInventory = initialInventory;
let finalCost = 0;
let currentCost = 1;
for (let i = 0; i < buy; i++) {
const currentInventorySizeString = (i + initialInventory).toString();
const isMod10 = currentInventorySizeString[currentInventorySizeString.length - 1] === '0';
finalInventory += 1;
if (isMod10) {
currentCost++;
}
finalCost += currentCost;
}
return [finalCost, finalInventory];
}
averages still the same, ~170 for u, ~400 for me
lyric mountain
```js
const getCost = ({initialInventory, buy}) => {
let finalInventory = i...
must return the cost
only the cost
nothing more than the cost
lyric mountain
it wont evaluate to true otherwise as it breaks the constraints
proven lantern
it's still wrong too
queen needle
Why isMod10 instead of n % 10
proven lantern
Why isMod10 instead of `n % 10`
probably faster to see if something is a '0' than doing math
perhaps
earnest phoenix
t
lyric mountain
lemme see if u understood the context
maybe ur thinking it's supposed to do something else
what are u trying to do?
proven lantern
they are buying buy items. those items change price depening on how many you already own. for every 10 you buy or own, the cost goes up by 1
you have an initial inventory and are buying more
lyric mountain
yes
quartz kindle
the initial price is also based on how many you own
proven lantern
const getCost = ({initialInventory, buy}) => {
let finalInventory = initialInventory;
let finalCost = 0;
let currentCost = 1;
for (let i = 0; i < buy; i++) {
const currentInventorySizeString = (i + initialInventory).toString();
const isMod10 = currentInventorySizeString[currentInventorySizeString.length - 1] === '0';
finalInventory += 1;
finalCost += currentCost;
if (isMod10) {
currentCost++;
}
}
return finalCost;
}
this should work
quartz kindle
if you already own 32, the initial price already starts at 4
quartz kindle
its 1268500 tho
lyric mountain
damn no matter what I do I cant get below 400 on java lmao
tf did u do up there tim
queen needle
I want to try but I don't have my laptop :(
lyric mountain
online editor
quartz kindle
tf did u do up there tim
recursion is not free :)
lyric mountain
tried removing recursion, little to no impact
quartz kindle
also less loops
lyric mountain
java most likely inlined it after the warmup phase
queen needle
There has to be a formula to do it without loops
queen needle
if you already own 32, the initial price already starts at 4
How would the initial price be 4 if it's +1 for every 10?
lyric mountain
also less loops
has to be the loop count, trynna find a way to reduce it atm
lyric mountain
How would the initial price be 4 if it's `+1` for every 10?
at 30 ur at 4 per unit
queen needle
Does it start at 1?
lyric mountain
1 - 10 = 1
11 - 20 = 2
21 - 30 = 3
31 - 40 = 4
queen needle
Ohhh
So you have the amount you're buying and what you start with, for 32 it's 4 as the initial cost, then for every 10 initial cost is increased by 1?
lyric mountain
no no, for every cumulative 10
so if u start at 32, it'll be 4 for the next 8 units
after that it'll be 5 per unit
limpid rain
Hello all
quartz kindle
it was closer in js
proven lantern
ok, it works for real this time
const getCost = ({initialInventory, buy}) => {
let finalCost = 0;
let currentCost = Math.floor(initialInventory / 10) + 1;
for (let i = 1; i <= buy; i++) {
const currentInventorySizeString = (i + initialInventory).toString();
const isMod10 = currentInventorySizeString[currentInventorySizeString.length - 1] === '0';
finalCost += currentCost;
if (isMod10) {
currentCost++;
}
}
return finalCost;
}
speed test time
i got this
queen needle
If I write one does anyone have anyway to benchmark it using a website or something?
lyric mountain
tim's testing on browser, I'm testing on intellij
quartz kindle
you can use jsben.ch or tio.run, but i like running everything on my browser dev tools first
and run it multiple times to get a sense of how stable it is
lyric mountain
aight it'll take a while
proven lantern
quartz kindle
also, whoever wants to test, look up chat historu can copy our functions too, so you test all of them in the same environment and compare
proven lantern
i want my faster cpu to give me an edge
frosty gale
y'all still at it 💀
quartz kindle
we're doing 1 million iterations
lyric mountain
quartz kindle
lmao
lyric mountain
the test btw
lyric mountain
yes, string ops are slow as hell
proven lantern
quartz kindle
the test btw
you dont need to do that comparison inside the loop lll
lyric mountain
you dont need to do that comparison inside the loop lll
ik, but it shouldn't matter much if I'm doing with all test cases
wanna make sure the code is cosmic-ray safe
quartz kindle
lmaoooo
proven lantern
i'll make a map of all the numbers that are mod 10 === 0 and use that instead
frosty gale
the test btw
very cheeky of you using tims loop to absorb the cold start performance 😳
lyric mountain
very cheeky of you using tims loop to absorb the cold start performance 😳
actually, just noticed ben's failed every warmup
forgot to convert to obj param
frosty gale
im going to do a battleless in a minute and start gaslighting
lyric mountain
make the params not an object please ben
proven lantern
btw name it getCostBen
here's the current state ```js
const getCostBen = (initialInventory, buy) => {
let finalCost = 0;
let currentCost = Math.floor(initialInventory / 10) + 1;
for (let i = 1; i <= buy; i++) {
const isMod10 = (i + initialInventory) % 10 === 0;
finalCost += currentCost;
if (isMod10) {
currentCost++;
}
}
return finalCost;
}
easier for me to plug the code
proven lantern
mmmmmkay
proven lantern
removing the isMod10 seems to help
const getCostBen = (initialInventory, buy) => {
let finalCost = 0;
let currentCost = Math.floor(initialInventory / 10) + 1;
for (let i = 1; i <= buy; i++) {
finalCost += currentCost;
if ((i + initialInventory) % 10 === 0) {
currentCost++;
}
}
return finalCost;
}```
lyric mountain
the code converted to java btw
proven lantern
it's not much faster
frosty gale
i wonder what the performance in c/c++ would be for all of them
if the compiler would optimise them all in a similar fashion
quartz kindle
do it
frosty gale
resulting in similar times
lyric mountain
it'll be very close to java
quartz kindle
take all our funcions and turn them indo a cpp benchmark
lyric mountain
static int getPrice(int has, int wants) {
final int basePrice = 1;
final int step = 10;
final int first = Math.min(step - has % step, wants);
final int second = wants - first - (wants - first) % step;
final int third = wants - first - second;
int out = (int) (first * (basePrice * Math.ceil(has / (double) step)));
if (second != 0) {
final int iterations = second / step;
final int multiplier = basePrice * ((has + first) / step) + 1;
for (int i = 0; i < iterations; i++) {
final int current = multiplier + i;
out += current * step;
}
}
if (third != 0) {
final int current = basePrice * ((has + first + second) / step) + 1;
out += third * current;
}
return out;
}
static int getCost(int initial, int buy) {
buy += initial;
int i = 0;
int cost = 0;
while (buy >= 10) {
cost += 10 * ++i;
buy -= 10;
}
if (initial == 0) {
return cost + buy * ++i;
} else {
return cost - getCost(0, initial) + buy * ++i;
}
}
also bens
lyric mountain
static int getCostBen(int initialInventory, int buy) {
int finalCost = 0;
int currentCost = (int) (Math.floor(initialInventory / 10d) + 1);
for (int i = 1; i <= buy; i++) {
final boolean isMod10 = (i + initialInventory) % 10 == 0;
finalCost += currentCost;
if (isMod10) {
currentCost++;
}
}
return finalCost;
}
this one will take a while to complete tho
frosty gale
ty
lyric mountain
long t = System.currentTimeMillis();
for (int i = 0; i < 999999; i++) {
if (getPrice(32, 5000) != 1268500) {
System.out.println("Tim's did not evaluate to 1.268.500");
break;
}
}
final long tim = System.currentTimeMillis() - t;
to test
queen needle
I'm driving myself crazy trying to get this in a one line formula
quartz kindle
good luck lol
quartz kindle
lmao
proven lantern
const divisibleByTenMap = {};
for (let i = 10; i < 10000000; i += 10) {
divisibleByTenMap[i] = true;
}
const getCostBen = (initialInventory, buy) => {
let currentCost = Math.floor(initialInventory / 10) + 1;
let finalCost = 0;
for (let i = 1; i <= buy; i++) {
finalCost += currentCost;
if (divisibleByTenMap[(i + initialInventory)]) {
currentCost++;
}
}
return finalCost;
}```ignore the startup time
queen needle
💀
proven lantern
it doesn't work either
lyric mountain
now let's test with random values
proven lantern
are you using my new and improved code?
lyric mountain
nope, u said it doesnt work
proven lantern
it does, i was just using the wrong signature
i was tricked
in js you usually use a spec object for params
lyric mountain
doing warmup
but it's likely much slower than ur former code, as it's still in warmup 2
proven lantern
for (let i = 10; i < 10000000; i += 10) {
divisibleByTenMap[i] = true;
}``` is initializing codeso it'll take a long time to start
but that builds a map that my function references
oh, but counting by 10 could be a trick
dont take my idea
lyric mountain
@quartz kindle still does well with random values (same seed)
js test with ben code is still warming up
proven lantern
lol
lyric mountain
just finished
proven lantern
i have a new plan to win
queen needle
How does mine return 1268504???
lyric mountain
off by 4
lyric mountain
<@180112943612952577> still does well with random values (same seed)
I'm honestly impressed with java's Random performance
it had little to no impact on the times
quartz kindle
js's Math.random() is very slow
lyric mountain
didn't test with random on js, used java cuz I can restart it with the same seed
to guarantee the tests are the same yet random
queen needle
How do I test in browser?
queen needle
I have a laptop now
lyric mountain
then F12
queen needle
I meant like benchmark, what is the boilerplate
lyric mountain
let t = performance.now();
for (let i = 0; i < 999999; i++) {
if (getPrice(32, 5000) !== 1268500) {
console.log("Tim's did not evalue to 1.268.500")
break;
}
}
const tim = performance.now() - t;
replace tim with ur name
quartz kindle
you dont need the inner if but yeah
lyric mountain
send the code too
proven lantern
const getCostBen = (initialInventory, buy) => {
let inventoryCountTracker = initialInventory;
let costTracker = 0;
// get inventory to a multiple of 10
let currentCost = Math.floor(inventoryCountTracker / 10) + 1;
let i = 0;
while (inventoryCountTracker % 10 !== 0) {
costTracker += currentCost;
inventoryCountTracker++;
i++;
}
currentCost++;
for (i; i < buy - 10; i += 10) {
costTracker += (currentCost * 10);
inventoryCountTracker += 10;
currentCost++;
}
for (i; i < buy; i++) {
costTracker += currentCost;
inventoryCountTracker++;
}
return costTracker;
}```C O S M I C 🌠 R A Y S
quartz kindle
okok xd
lyric mountain
warmup is faster this time
lyric mountain
quite impressed
quartz kindle
oh my
proven lantern
i can remove the inventory counting to make it faster
const getCostBen = (initialInventory, buy) => {
let inventoryCountTracker = initialInventory;
let costTracker = 0;
// get inventory to a multiple of 10
let currentCost = Math.floor(inventoryCountTracker / 10) + 1;
let i = 0;
while (inventoryCountTracker % 10 !== 0) {
costTracker += currentCost;
inventoryCountTracker++;
i++;
}
currentCost++;
for (i; i < buy - 10; i += 10) {
costTracker += (currentCost * 10);
currentCost++;
}
for (i; i < buy; i++) {
costTracker += currentCost;
}
return costTracker;
}```in my case I just need to somehow reduce the loop iterations
proven lantern
in my case I just need to somehow reduce the loop iterations
my code recudes by 10x
unless you do that too
lyric mountain
lyric mountain
my code recudes by 10x
yep
function getCost(initial, buy) {
buy += initial;
let i = 0;
let cost = 0;
while (buy >= 10) {
cost += 10 * ++i;
buy -= 10;
}
if (initial === 0) {
return cost + buy * ++i;
} else {
return cost - getCost(0, initial) + buy * ++i;
}
}
actually
proven lantern
const getCostBen = (initialInventory, buy) => {
let inventoryCountTracker = initialInventory;
let costTracker = 0;
// get inventory to a multiple of 10
let currentCost = Math.floor(inventoryCountTracker / 10) + 1;
let i = 0;
while (inventoryCountTracker % 10 !== 0) {
costTracker += currentCost;
inventoryCountTracker++;
i++;
}
currentCost++;
const buyMinusTen = buy - 10;
for (i; i < buyMinusTen; i += 10) {
costTracker += (currentCost * 10);
currentCost++;
}
for (i; i < buy; i++) {
costTracker += currentCost;
}
return costTracker;
}```i made this change js const buyMinusTen = buy - 10; for (i; i < buyMinusTen; i += 10) { costTracker += (currentCost * 10); currentCost++; }
should be faster than doing the math ever time
it should beat tim now
ezpz
quartz kindle
wat
queen needle
I gave up lol
proven lantern
ignore mine, I'm doing some changes
can you make tim's slower so i win?
quartz kindle
lmao
proven lantern
pls
proven lantern
ignore mine, I'm doing some changes
changes are looking good
lyric mountain
trynna remove multiplication altogether
proven lantern
this is 3x slower for me
maybe because it assigns
i'll do it all in one statement
this is just a bit slower than x * 10
queen needle
Lmao
lyric mountain
VOILA! I did it
function getCost(initial, buy) {
let i = initial / 10 + 1 | 0;
let cost = 0;
let step = 10 - initial % 10;
while (buy >= 10) {
cost += step * i++;
buy -= step;
step = 10;
}
return cost + (initial / i | 0) - (initial % 10) + buy * ++i;
}
const getCostBen = (initialInventory, buy) => {
let inventoryCountTracker = initialInventory;
let costTracker = 0;
// Get inventory to a multiple of 10
let remainingItems = inventoryCountTracker % 10;
if (remainingItems !== 0) {
costTracker += Math.floor(inventoryCountTracker / 10) + 1;
inventoryCountTracker += 10 - remainingItems;
}
// Calculate cost for buying in multiples of 10
const buyMinusTen = buy - 10;
let currentCost = Math.floor(inventoryCountTracker / 10) + 1;
let i = inventoryCountTracker;
for (; i < buyMinusTen; i += 10) {
costTracker += currentCost * 10;
currentCost++;
}
// Calculate cost for buying the remaining items
let remainingBuy = buy - i;
costTracker += currentCost * remainingBuy;
return costTracker;
};
ai helped with this one
does ai win?
ai + ben
quartz kindle
```js
function getCost(initial, buy) {
let i = initial / 10 + 1 | 0;
let...
nice :)
quartz kindle
lmao
proven lantern
and the result is wrong
i like how ai just lies
the result is actually 1252404
quartz kindle
xD
proven lantern
i can't get chatgpt to give me anything that works
queen needle
Is there anyway to do it without a loop? I doubt it but if there was wouldn't that be the fastest
lyric mountain
nice :)
This taught me to use | 0 on any calcs where I don't care abt decimals
It's the equivalent of (int) value in java
quartz kindle
ye, it forces conversion into int32
lyric mountain
Tbh I hate how any division sets the number to double
quartz kindle
isnt a double required to make divisions in the cpu?
lyric mountain
I think it's only required if you want the decimals
Else it gets floored as there are no bytes for decimals
lyric mountain
const buyMinusTen = buy - 10; YOU DON'T SAY
proven lantern
`const buyMinusTen = buy - 10;` YOU DON'T SAY <:KEKW:727825308404940800>
i like my variable names to make sense
earnest phoenix
```js
function getCost(initial, buy) {
let i = initial / 10 + 1 | 0;
let...
Good job, that's a pretty clever way to do that
Implementing that in C++ with optimization level 3 takes 1ms for 1 million iterations
lyric mountain
Ty, that was one hell of a trip
lyric mountain
Implementing that in C++ with optimization level 3 takes 1ms for 1 million itera...
Ah ffs
tho...are u trying with random or fixed values?
cuz it's probably compiling the function to a constant as the params never change
earnest phoenix
Completely disabling optimizations makes it take 3 seconds
With completely random values (with optimization level 3) it takes 109ms for 1 million iterations
queen needle
Now do that in js
lyric mountain
we...kinda did
earnest phoenix
#include <chrono>
#include <cstdlib>
#include <iostream>
size_t get_price(size_t has, size_t wants) {
size_t cost{0};
size_t cost_multiplier{has / 10 + 1 | 0};
size_t step{10 - has % 10};
while (wants >= 10) {
cost += step * cost_multiplier++;
wants -= step;
step = 10;
}
const size_t cost_mult{has / cost_multiplier | 0};
return cost + cost_mult - (has % 10) + wants * ++cost_multiplier;
}
int main() {
std::srand(std::time(nullptr));
auto start = std::chrono::high_resolution_clock::now();
for (size_t i = 0; i < 999'999; i++) {
[[maybe_unused]] volatile auto foo = get_price(rand(), rand());
}
auto end = std::chrono::high_resolution_clock::now();
std::cout << std::chrono::duration_cast<std::chrono::milliseconds>(end -
start)
.count()
<< "ms\n";
}
so about 4x faster than our js solutions
queen needle
we...kinda did
I meant the times not the code
lyric mountain
and 1.2x faster than java
considering tim's code ofc
didn't convert my solution to java to see
earnest phoenix
Maybe I should try converting Tim's code to C++ as well, see how it compares to yours in C++
Where's Tim's latest one?
lyric mountain
```java
static int getPrice(int has, int wants) {
final int basePrice = 1;
...
^
he didn't change it since this message
it's the getPrice btw
earnest phoenix
Alright let me see
lyric mountain
there's a version I converted to java, if it's easier for u to work with nvm, that is the java version
the original js code is a little higher up
earnest phoenix
~~there's a version I converted to java, if it's easier for u to work with~~ nvm...
Does final mean it's constant?
lyric mountain
yep
lyric mountain
ohno
quartz kindle
lyric mountain
getPrice7 = (a, b) => true
quartz kindle
it might have bugs
but i tested a bunch of different values compared to getPrice and they all return the same results
but this is a pure math way, zero loops
function getPrice7(has, wants) {
const hasrem = has % 10;
const hasbase = has - hasrem;
const hasvalue = hasbase + hasbase * (0.5 * ((hasbase / 10) - 1)) + hasrem * (Math.ceil(hasbase / 10) + 1);
const total = has + wants;
const r = total % 10;
const x = total - r;
return (x + x * (0.5 * ((x / 10) - 1)) + r * Math.ceil(total / 10)) - hasvalue;
}
ah nvm i found some values that the result is wrong
more specifically if has is a multiple of 10, the final result has 10 too much
should be easy to fix
lyric mountain
mf solved the math
quartz kindle
nvm its not easy to fix, still looking at where the extra 10 is coming from lmao
queen needle
I knew someone was going to figure It out
quartz kindle
i'll figure it out later g2g do something else now
earnest phoenix
@lyric mountain @quartz kindle
proven lantern
const getCostBen = (initialInventory, buy) => {
let inventoryCountTracker = initialInventory;
let costTracker = 0;
let currentCost = Math.floor(inventoryCountTracker / 10) + 1;
if (inventoryCountTracker % 10 !== 0) {
const remainingCount = 10 - (inventoryCountTracker % 10);
costTracker += currentCost * remainingCount;
inventoryCountTracker += remainingCount;
currentCost++;
}
const finalInventorySize = initialInventory + buy;
const finalInventorySizeMinus10 = finalInventorySize - 10;
for (; inventoryCountTracker < finalInventorySizeMinus10; inventoryCountTracker += 10) {
costTracker += currentCost * 10;
currentCost++;
}
const remainingCount = finalInventorySize - inventoryCountTracker;
costTracker += currentCost * remainingCount;
return costTracker;
};
i fixed the ai code
queen needle
When Tim fixes it you'll have to test his new one
proven lantern
it's not much faster
lyric mountain
<@350836145921327115> <@180112943612952577> <:KEKW:727825308404940800>
damn
earnest phoenix
KuuHaKu doesn't even use C++
proven lantern
lol
lyric mountain
tim and I are entangled, same time for both algos lmao
proven lantern
did my last one win?
proven lantern
```js
const getCostBen = (initialInventory, buy) => {
let inventoryCountTrac...
fastest one
lyric mountain
did my last one win?
idk, not with my testing environment anymore
proven lantern
i'll find them
lyric mountain
but it's simple, just declare the functions, run 100 loops of 999999 for each function to warm then and then measure how long it takes for them to run 999999 times
proven lantern
i win sometimes!
lyric mountain
haka 💀
earnest phoenix
When Tim fixes it you'll have to test his new one
Of course
tulip ledge
@quartz kindle @lyric mountain (sorry for ping) I found a formula. I was talking to a friend of mine who's also in my math class and we came up with a formula and it goes as follows:
Let's assume the player starts with 0 items and wants to buy 10k + r items with k being a whole multiple and r being a natural number.
10 items at 1$, 10 items at 2$, ..., 10 items at k$, r items at (k + 1)$
Total cost is 10(1+2+...+k) + r(k+1) = 10k(k+1)/2 + r(k+1) = 5k(k+1) + r(k+1) = (5k+r)(k+1)
let's say we buy N items then k = floor(N / 10) and r = N % 10
And we start at M items instead of 0
We get cost of buying M + N from 0 - cost of buying M from 0
thus we get cost = (5 * floor((M + N) / 10) + (M + N) % 10) * (floor((M + N) / 10) + 1) - (5 * floor(M / 10) + M % 10) * (floor(M / 10) + 1)
filling in 6 as M and N as 10 gives the correct answer (16) and filling in 6 as M and N as 48 gives us 168, which is also correct. This equation holds true for all natural numbers
earnest phoenix
```js
const getCostBen = (initialInventory, buy) => {
let inventoryCountTrac...
Your solution is so slow that it would take an entire day for it to complete in 1 million iterations, just tested in C++
lyric mountain
lmao
proven lantern
i feel like we are being trolled
earnest phoenix
Probably, though I might be converting something of their solution incorrectly to C++, here's what I did:
// Ben's solution
size_t get_price_ben(size_t has, size_t wants) {
size_t inventory_count_tracker{has};
size_t cost_tracker{0};
size_t current_cost{
static_cast<size_t>(floor(inventory_count_tracker / 10) + 1)};
lcpp_x:
if (inventory_count_tracker % 10 != 0) {
const size_t remaining_count{10 - (inventory_count_tracker % 10)};
cost_tracker += current_cost * remaining_count;
inventory_count_tracker += remaining_count;
current_cost++;
}
const size_t final_inventory_size{has + wants};
const size_t final_inventory_size_minus_10{final_inventory_size - 10};
for (; inventory_count_tracker < final_inventory_size_minus_10;
inventory_count_tracker += 10) {
cost_tracker += current_cost * 10;
current_cost++;
}
const size_t remaining_count{final_inventory_size - inventory_count_tracker};
cost_tracker += current_cost * remaining_count;
return cost_tracker;
}
From what I'm seeing it's 1:1 to their JavaScript version
earnest phoenix
i feel like we are being trolled
You're probably not benchmarking your solution with different values and are just using the exact same values which might explain why you're getting a fast output in JavaScript
This is how I'm benchmarking them, with completely different values
https://sourceb.in/2lIKbpa0UU
quartz kindle
mf solved the math
the function is correct, the previous one we have been using is the one is wrong lmao
for example
if i have 19, and want to buy 2
the first one should cost 2, the second should cost 3, correct?
total 5
the previous function was giving the wrong resut
lyric mountain
what in the fuck are up with those names
earnest phoenix
You're going quite high with both the parameters, have you tried with small values to see if your solution has a regression?
proven lantern
what in the fuck are up with those names
thanks
quartz kindle
what in the fuck are up with those names
check this out, the old function was totally wrong
lyric mountain
check this out, the old function was totally wrong
I'll try it later
earnest phoenix
@proven lantern these are the kind of values passed to your function in my benchmarks, I think it's pretty clear why your solution takes so incredibly long
Your solution tries to loop that amount of times combined
earnest phoenix
You shouldn't be using dates to measure performance in JavaScript, use performance.now()
quartz kindle
<@180112943612952577> <@350836145921327115> (sorry for ping) I found a formula. ...
i figured out a similar solution, but your's is more elegant
congrats
earnest phoenix
There's also console.time() and console.timeEnd()
radiant kraken
Probably, though I might be converting something of their solution incorrectly t...
convert it to Rust
now it's blazingly fast
quartz kindle
I'll try it later
so:
// tim's new solution
function getPrice7(has, wants) {
const hasrem = has % 10;
const hasbase = has - hasrem;
const hasvalue = hasbase + hasbase * (0.5 * ((hasbase / 10) - 1)) + hasrem * (Math.ceil(hasbase / 10) + 1);
const total = has + wants;
const r = total % 10;
const x = total - r;
return (x + x * (0.5 * ((x / 10) - 1)) + r * Math.ceil(total / 10)) - hasvalue;
}
// vigintille's friend's math solution
function test2(M, N) {
return (5 * Math.floor((M + N) / 10) + (M + N) % 10) * (Math.floor((M + N) / 10) + 1) - (5 * Math.floor(M / 10) + M % 10) * (Math.floor(M / 10) + 1)
}
Oh my god
radiant kraken
hi pancake
earnest phoenix
now it's blazingly fast
Rust blazingly fast!!!!!!!! 💯💯💯💯🔥🔥🔥🔥
radiant kraken
long time no see
queen needle
I had a nearly identical solution to the second one, I just multiplied by something wrong
radiant kraken
@queen needle your website is awesome btw
i love it when portfolios are actually creative
and not steal the same common traits from others
quartz kindle
its empty for me, is it supposed to have something?
just has a background, a plus button and a zoom button
queen needle
You draw the windows by clicking and dragging to a different point, the bigger the better
radiant kraken
just has a background, a plus button and a zoom button
try drawing a rectangle with your mouse
quartz kindle
ahh
queen needle
Oh yeah, I meant to when I made it but I never did 💀
radiant kraken
@civic scroll what do you think of pancake's website
quartz kindle
ok this is actually amazing
a couple suggestions tho, add a way to resize windows :3
radiant kraken
^
quartz kindle
also are the links in the projects page supposed to work?
queen needle
So okay
They are supposed to open another window, but I did the math for centering it wrong and never fixed it
quartz kindle
ah maybe its gtting blocked by some ad blocker
radiant kraken
i dont think so @quartz kindle
queen needle
Open like another of the windows on the site, not in another tab
I also think of the urls got changed
quartz kindle
nope still doesnt work
ah i see
anyways this is the most creative thing i've seen in a long time
congrats!
queen needle
I love the portfolio, but idk anyway to make it work for mobile, any suggestions would be nice because I don't want to change it, but if I can't get it to work for mobile I might have to
quartz kindle
tho from a UX perspective its a bit funky, design and creativity wise its godlike
queen needle
tho from a UX perspective its a bit funky, design and creativity wise its godlik...
I wanted to make like one of the tutorials where it tells you to click things and stuff, I just never did, like an "onboarding" feature
quartz kindle
haha coool
queen needle
I'd love any ideas of how to adapt it to mobile
earnest phoenix
so:
```js
// tim's new solution
function getPrice7(has, wants) {
const hasr...
Your new solution gives 1268498 for 32, 5000
radiant kraken
I'd love any ideas of how to adapt it to mobile
just implement the same thing in mobile?
queen needle
Windows wouldn't really be big enough imo
lyric mountain
Your new solution gives `1268498` for `32, 5000`
we need someone to calc the actual value for that so we have something to compare against
quartz kindle
Your new solution gives `1268498` for `32, 5000`
?
radiant kraken
Windows wouldn't really be big enough imo
make it smaller
or make it a mask so users can just scroll through
queen needle
I have an idea for a new one where the windows are already made and you can go around to different sections to view them
earnest phoenix
?
This is how I've converted yours to C++
// Tim's second solution
size_t get_price_tim2(size_t has, size_t wants) {
const size_t hasrem{has % 10};
const size_t hasbase{has - hasrem};
const size_t hasvalue{
static_cast<size_t>(hasbase + hasbase * (0.5 * ((hasbase / 10) - 1)) +
hasrem * (std::ceil(hasbase / 10) + 1))};
const size_t total{has + wants};
const size_t r{total % 10};
const size_t x{total - r};
return (x + x * (0.5 * ((x / 10) - 1)) + r * std::ceil(total / 10)) -
hasvalue;
}
im doing multiplication with floats, shouldnt you be using double instead of size_t
well you do have a cast
idk then
earnest phoenix
Oh, is hasvalue supposed to be a double?
quartz kindle
```js
function getCost(initial, buy) {
let i = initial / 10 + 1 | 0;
let...
btw this is completely wrong lel
radiant kraken
JS numbers are doubles behind the scenes so
