#Legion 5 15ACH6H 3070
1 messages · Page 2 of 1
Yeah only just made the link now 🤦
I saw there was a difference
But didn't think it was importsnt
Ok I put the ram in
Its detected under 0x50
sudo ./spd_dump 2 0x50 something
I write something yeah?
doesnt matter 🙂
Name.bin i asume
I didn't put .bin yh
it doesnt need to have an extension
Is that not just going to compare the good stick with itself? Not the mod stick we flashed
we can check that too
No yeah I'm a fucking moron
You're right
with an untested chipset i dont know how much you can go
you can still experiment with it if you feeling tech savy
I think he needs ro edit timings manually to match the thaiphoon screen
did you diff them?
if they are same
Maybe my dump gives some other info why the stick wont work anymore but the timings can work
No reply for diff
reboot time then
Pray for me
sure
nice to hear
Thanks for the help bruh you saved me fr @serene grove @molten rover
Cihan, how did you edit the timings on your dump to match my thaiphoon screen?
i used smokeless dumper
opened it on ddr4xmpeditor on windows
edited it saved it and flashed useing smokeless spd_write tool on linux
Seems not, but you can do what cihan mentioned snd edit the timings manually in the xmpeditor
Then flash back and test
You now know all the commands to flash in Linux👀
Like a Linux GURU
i'll need to do some research and figure out what each thing does
what should i not touch in ddr4xmpeditor?
I dont know that, i havent used that myself
well for thaiphoon then7
For thaiphoon you can copy my values, everything that does nothing is still at stock values👀
I have a little self-created list on my work laptop about some values
But im not close to that machine now
no worries
i might try throwing all your values into xmpeditor myself and seeing if it works
or maybe verse myself with hex editor and figure out how to change min cas latency myself
@molten rover how much of a difference did tweaking min cas latency even make?
%1?
what would you say IS low?
rightttt
I dont know what you mean with this
I just bought thaiphoon and changed all settings 1 by 1
if it ran then great
Took me like a entire day to get to my cl18 vakues
fair enough
No idea what every setting does
haha
ive spent 2 days working to find better stabler values
how long are you meant to test the ram for after a tweak? 2 hours?
i did 2 hour runs
maybe 48 hours?
could be
you test many different timings and to decide they need to pass stability test
alright
In this guide it’s explained everything about timings
thanks
hmm i might buy thaiphoon then, it'd be much easier on windows especially if i have to put the computer on its side
cant get bluetooth working on linux so i have to use the keyboard and shi sideways
but £23.43 is a bit steep im gna have to think about it
I did 12.2k on 140w
And that was with not going outside, just iets home during autumn
yo @serene grove i figured out how to edit min cas latency time
its a bit complicated but it takes 5 minutes to understand and ive confirmed it works by checking the new spd in thaiphoon burner free edition. actually im not sure it would work. looking in thaiphoon it says spd checksum failed on the spds where i tweak min cas latency. will have to flash it sometime and see
hmm on second look something really weird is going on
looks like shortcuts in thaiphoon when using .spd files are bugged
I read a question towards me this morning when i woke up but cant see it anymore
Is it not needed anymore?
Yeah I realised you sent your whole thaiphoon file so I just checked that lol
OK, or in your words Calm

sorry jumping in chat late, i got 10633 score on stock 2x16gb, and then upgraded to dual rank kingston fury and got 10770 score. i upgraded to the recommended ram from youtube videos like jarrods tech etc it was hard to find them bit dissapointed in the increased score lol im on windows 11 though, no background apps
It looks like your rams was already good, as anything close to 11k on 5800h is good
Yeah I had a feeling of that after I upgrade, which is sad because all the forums said there was a thing where legion 5s and 7s were coming with cheap slow single rank ram
If I follow the above comments and advice and push the score to 12k is that going to be visible when gaming stable fps increase ?
Your stock rams must be for some reason better, you can easily refund online orders as unwanted
Pushing it using info above is for benchmarks mostly, no way my benchmark overclock will run any game without instant crashing
But if you do scroll around a bit above you'll see some info relevant to games and yeah that'll boost in game performance by a bit
If you're on legion 5 regular 2021 3070 then flashing a 140w clevo vbios is the first step to better performance
Ask around for help on it
Reading a few messages from around here is also relevant for game performance
But you need to test overclocks using timespy to get your most stable overclock (most you can push oc numbers with no artifacts whatsoever)
Not sure what you mean
i was just replying to your comment so i could use it as a sort of link to get to to your comment easily
What was your last power run again, TS link i mean
best one?
Was that inside or with extra cold air?
window
How the frock do u get such coool temps
Whats your best one without extra cold air?
did it infront of a window at 6 degrees
celcius
not sure tbh
no link but that was at ambient
1rx8?
run it infront of an ac
nah 2rx8
Yeh that's how I get cool temps
yeah thats good
temp in india is crazy bro i had to sit infront of a fan or ac all trip
Do a max OC run in normal ambient and sent link
That will show a good startingpoint
Haha
https://www.3dmark.com/spy/33780522 found the link
Mine at 18c
Ok, thats on 150w i assume
140w
Awesome score
uk
My coolest on 18c was 70c
true
Ok, your average speed is 1782 so you should go a little bit lower.
- Your core that it gets to 1755 at .837
Awesome, so for game stable you do a +105 core and +500 mem
Then the cutoff at .837
Im expecting that to be game stable without artifacting
Make that curve and test in SOTTR benchmark
If stable in SOTTR then its like stable everywhere👀
for now ive been running +150 (no curve) with no artifacts or crashing in games for a week now
should i still drop to 105?
or does cutoff curve affect stability in some way?
If you are OK with that then go ahead
Im expecting what i suggested to be better then stock score but cooler temps
But if what you mentioned is stable and you are pleased with perf/temps just use that
Cutoff limits voltage to that point , so should yield better temps
And same perf
i'll try both then, thanks
Nice thing of SOTTR testing, it gives a bit better visual of actuall FPS profit compared to TS
Mine is a legion 7 5800h, 3070, 2x16gb fury 2rx8. Thankyou I'll look through the comments more, I did notice the bios from clevo, I have seen that before in different forums for different laptops wasn't sure if it was stable though, I haven't researched enough how using different laptop bios works