AMD Mem Tweak XL - Read/modify timings/pp/straps on the fly

[scryptr]

DEAD CARD DETECTION--

My Vega64 rig was stable for 2 days mining CNr on NiceHash with TeamRedMiner (2.2kH/s per GPU), but I noticed that poolside hashrate varied a lot, both on the NiceHash dashboard (wild swings +/-), and on the TRM console (more steady but low).  I thought to experiment with ETH.

Switching to ETH from CNr, TRM to PhoenixMiner, I initially achieved 48MH/s ETH per card.  But, very shortly 1 GPU was reported dead and not hashing.  The rig generally began behaving badly and I eventually pulled that GPU.  I tried ETH overnight with the remaining 3 cards but woke to a zero hash rate.

Unplugging and replacing the "dead" GPU a few times brought no better results, and switching back to TRM mining CNr still wound up with a dead GPU detected and failed launches.

So, I commented out my call to AMDmemtweak in my "custom.sh" file (a script file run at boot by ethOS to launch utilities and such).  After that I configured the rig to mine VertCoin (lyra2v3) with TRM.  Lyra2v3 is a compute heavy algorithm, and I have my memory clocked low (500mhz).  TRM booted just fine, both without and with my "undead" GPU.  The cards are hashing VTC at about 100MH/s each.

I sincerely expected to RMA the card last night.  Today it works fine mining Lyra2v3, but at a higher wattage (250+W/GPU) than when mining CNr (200W/GPU) and ETH (150W/GPU).

I need to do some more reading about the strap values, I think.  Any tips or comments left on the matter will be appreciated.       --scryptr

[pbfarmer]

DEAD CARD DETECTION--

My Vega64 rig was stable for 2 days mining CNr on NiceHash with TeamRedMiner (2.2kH/s per GPU), but I noticed that poolside hashrate varied a lot, both on the NiceHash dashboard (wild swings +/-), and on the TRM console (more steady but low).  I thought to experiment with ETH.

Switching to ETH from CNr, TRM to PhoenixMiner, I initially achieved 48MH/s ETH per card.  But, very shortly 1 GPU was reported dead and not hashing.  The rig generally began behaving badly and I eventually pulled that GPU.  I tried ETH overnight with the remaining 3 cards but woke to a zero hash rate.

Unplugging and replacing the "dead" GPU a few times brought no better results, and switching back to TRM mining CNr still wound up with a dead GPU detected and failed launches.

So, I commented out my call to AMDmemtweak in my "custom.sh" file (a script file run at boot by ethOS to launch utilities and such).  After that I configured the rig to mine VertCoin (lyra2v3) with TRM.  Lyra2v3 is a compute heavy algorithm, and I have my memory clocked low (500mhz).  TRM booted just fine, both without and with my "undead" GPU.  The cards are hashing VTC at about 100MH/s each.

I sincerely expected to RMA the card last night.  Today it works fine mining Lyra2v3, but at a higher wattage (250+W/GPU) than when mining CNr (200W/GPU) and ETH (150W/GPU).

I need to do some more reading about the strap values, I think.  Any tips or comments left on the matter will be appreciated.       --scryptr

If your cards aren't adequately cooled, a super high REF can cause hw errors.  My vegas @ 40 are fine w/ 15600.  My old, cranky, and hotter ref 56 doesn't seem to like that (it's temperamental even w/o timing mods.) That's what I'd look at first. After that, I've found RCDRD and RCDWR to be the most common culprits of hw errors.

As for poolside h/r variability - it is a fact of life w/ CN in my experience.  24hr h/r reports on the pool are helpful, but generally, those are not even long enough avgs (unless your diff is super low.)  Best bet is to look at miner reported pool rate, after at least a couple days, if you can keep your rig up that long.

[Mashy81]

Hey Mashy, sorry sir, no XMR address.
You can always use a shift service if you want to tip.
coinswap, changelly.. etc..

Get yourself one, I'll be happy to mine some XMR for you

Haha, ok ok, you convinced me ;-)

Check OP

And thank you!

Your xmr address comes up with a space when copied and pasted.

[SamAlackass]

Anyone else having this issue? I run the .exe with --current and it shows everything set at 0. Trying to apply any changes I get some confirmation but nothing really changes.

Obviously windows (10), testing with a hynix 56 on 18.5.1 and 18.6.1 with the same result.

Still looking for a solution to that. It just works on one system (win10, 18.6.1 driver) but not the other. I didn't even download AMDMemoryTweak again, just copied the entire folder over to the second system. Something obvious I may be missing?

[serg74]

Share settings for R9 390. Thank you.

[bobymoore]

For the noobs, could somebody explain the steps?
Exe don't open, do I have to create a bat file? Explain please....

[PIOUPIOU99]

I do not really use exchange, so I put my kh with your xmr on supportxmr for the day.
thx.

[saltednutz]

so this is where pbfarmer has been hiding

[Mashy81]

For the noobs, could somebody explain the steps?
Exe don't open, do I have to create a bat file? Explain please....

Yes, create a .bat file something like:

WinAMDTweak.exe --i 0,1,2 --rfc 248 --ref 15600     etc...

--i  is the gpus as per the bus order, if you are applying to all just leave it out. other commands are examples.

hope this helps

[Kerppa77]

I have now run ~72 hours w/o any issues with 24 Vega64 GPUs(good and bad ones) mixed in 4 separate rigs ... CNR with core 1407Mhz /860-970mV , Mem 1080-1100Mhz/860-970mV.

I would say this is pretty stable setup. Pool side 24h avg hashrate fluctuates in the range 54,1-54,4 kH/s

--CL 19 --RAS 28 --RCDRD 12 --RCDWR 5 --RC 44 --RP 12 --RRDS 3 --RRDL 3 --REF 15600 --RFC 250

I am happy and have in my mind... "dont touch when things go smoothly".

That’s an insane voltage range, and I can’t fathom how 970mv could possibly be necessary. Between 9 64s and a flashed 56, I run avg 834mv, none over 850, for similar clocks/timings.  That puts me @ 178w per GPU at the wall.

At 970, you’re running at least 35% higher power, meaning over 240w per card.

It sounds like you have some ppt issues or something, maybe you’re running your SOC @ 1200mhz unnecessarily (it should be 1107 @ your clocks.)  Although, even then, you prob shouldn’t need much more than 925mv.

Oops... I made a quick check with HWtool and yes... I really have SOC max @ 1199Mhz. I have used the Hellae 920mV PPT and then adjusted with Overdrive Ntool the Core P7 and Mem P3. I havent touch the SOC. After memory timings my rigs are pulling like 1420-1500w from the wall per rig. My rigs run like 150w w/o GPUs and after OverdriveNtool the wattage is ~250w before the miner starts and all the GPU led lights are fully on.
This sounds to me like 1270 to 1350w with 6 GPU and on avg. like 210 - 225w/GPU. :O

Thanks for pointing out the SOC. I need to drop to 1107.

When I tried to scribble down a new Vega tuning guide to include in the TRM 0.4.4 release, I was debating if I should have mentioned this or not. I opted not to, that text was long enough already anyway.

However, I know more than a few skilled and knowledgeable Vega tuners that ALL have been stuck with old PPTs from the fall of 2017 (brnsted, Hellae) that for some reason all have 1200 for SoC. At the time, I guess it was considered highly beneficial for stability. Fast forward a few years to times where you don't make $10/day/Vega and power draw is a crucial part of your setup, this trade-off no longer holds.

In the common Google sheet (https://docs.google.com/spreadsheets/d/1-rhYsaRXO1ahk3PyrEgT9gXzs7ImAzh-sbqtgwy8HQg) used for PPT tweaking you can edit the SOC clk value but guess what: it doesn't change any nrs in the PPT, the cells aren't connected(!!!). You have to change the SOC clk value in cell AE26, then manually copy cells AG-AI:26 to T-V:11. You can probably guess why I may sound a little pissed off  .

Yeah.... I fixed my file so that these cell are now connected. Will test today what kind of mV I can get with 1107Mhz SOC.

[cscheat]

Hey guys,

getting error message while trying to run --current

Failed to load EIO.dll

[iRonNuke]

Hey guys,

getting error message while trying to run --current

Failed to load EIO.dll

Try to reinstall MSI Afterburner/other OC programs. Do you have an Asus monitoring software, reinstall or uninstall that!

[rednoW]

Hey guys,

getting error message while trying to run --current

Failed to load EIO.dll

Are you using SRBminer? It blocks it ((

[JuanHungLo]

Hey guys,

getting error message while trying to run --current

Failed to load EIO.dll

Are you using SRBminer? It blocks it ((

It does on v 1.8.4, not on the prior versions.

[dragonmike]

Hey guys,

getting error message while trying to run --current

Failed to load EIO.dll

Are you using SRBminer? It blocks it ((

Hah. What a surprise.
SRB basically took eliovp's free tool and integrated a few self-compiled mem timing straps into his latest closed-source fee-including miner.

TRM had the decency not to integrate it. They chose to add a timing-optimised mode - should you want to run eliovp's tool whilst using TRM.

[doktor83]

Hey guys,

getting error message while trying to run --current

Failed to load EIO.dll

Are you using SRBminer? It blocks it ((

Hah. What a surprise.
SRB basically took eliovp's free tool and integrated a few self-compiled mem timing straps into his latest closed-source fee-including miner.

TRM had the decency not to integrate it. They chose to add a timing-optimised mode - should you want to run eliovp's tool whilst using TRM.

Yes i was waiting for this. Now if you look closely i couldn't integrate ElioVP's tool into my miner, but i guess you just don't have enough knowledge to inspect these kind of things.
Oh wait, here's one tip : we are not even using the same kernel driver to access phyisical memory, that means i can;t be using the ASUS files he uses because his complete implementation is built on that.
But how should you know that.. Claymore is coming with built in timings manipulation too.. Guess he took ElioVP's tool too.
Also what is a 'self-compiled mem timing straps' if you could explain, cause i never heard about compiled mem timing straps.

[doktor83]

Hey guys,

getting error message while trying to run --current

Failed to load EIO.dll

Are you using SRBminer? It blocks it ((

Simply use the --disabletweaking parameter and you can use ElioVP's tool with it.

[dragonmike]

Hey guys,

getting error message while trying to run --current

Failed to load EIO.dll

Are you using SRBminer? It blocks it ((

Hah. What a surprise.
SRB basically took eliovp's free tool and integrated a few self-compiled mem timing straps into his latest closed-source fee-including miner.

TRM had the decency not to integrate it. They chose to add a timing-optimised mode - should you want to run eliovp's tool whilst using TRM.

Yes i was waiting for this. Now if you look closely i couldn't integrate ElioVP's tool into my miner, but i guess you just don't have enough knowledge to inspect these kind of things.
Oh wait, here's one tip : we are not even using the same kernel driver to access phyisical memory, that means i can;t be using the ASUS files he uses because his complete implementation is built on that.
But how should you know that.. Claymore is coming with built in timings manipulation too.. Guess he took ElioVP's tool too.
Also what is a 'self-compiled mem timing straps' if you could explain, cause i never heard about compiled mem timing straps.

Whatever, Dok. You know what you did and didn't. So you play on semantics.
What you did (and Claymore is going to), is a direct consequence of eliovp releasing his open-source tool. You could have been giving him credit at least before putting it (in whatever form) into your paid miner.

[doktor83]

Hey guys,

getting error message while trying to run --current

Failed to load EIO.dll

Are you using SRBminer? It blocks it ((

Hah. What a surprise.
SRB basically took eliovp's free tool and integrated a few self-compiled mem timing straps into his latest closed-source fee-including miner.

TRM had the decency not to integrate it. They chose to add a timing-optimised mode - should you want to run eliovp's tool whilst using TRM.

Yes i was waiting for this. Now if you look closely i couldn't integrate ElioVP's tool into my miner, but i guess you just don't have enough knowledge to inspect these kind of things.
Oh wait, here's one tip : we are not even using the same kernel driver to access phyisical memory, that means i can;t be using the ASUS files he uses because his complete implementation is built on that.
But how should you know that.. Claymore is coming with built in timings manipulation too.. Guess he took ElioVP's tool too.
Also what is a 'self-compiled mem timing straps' if you could explain, cause i never heard about compiled mem timing straps.

Whatever, Dok. You know what you did and didn't. So you play on semantics.
What you did (and Claymore is going to), is a direct consequence of eliovp releasing his open-source tool. You could have been giving him credit at least before putting it (in whatever form) into your paid miner.

Yes, offsets from his code helped me finish an ongoing project. You wrote i took his tool, and emphasized how im using it in a closed source fee including miner.
Now it's whatever...

1. Look at the readme that comes with SRB
2. Look at the first page in SRB thread, bottom of 1st post
3. Look at ElioVP's btc donation address. 3 donations, Claymore, Me and 1 more.

Please next time before you write something like this do your homework first. But you wont, it's whatever..

[doktor83]

Whatever, Dok. You know what you did and didn't. So you play on semantics.
What you did (and Claymore is going to), is a direct consequence of eliovp releasing his open-source tool. You could have been giving him credit at least before putting it (in whatever form) into your paid miner.

+1 on this definitely...
[/quote]

yes and the direct consequence of your gpu's hashing faster it that ElioVp released his tool. So what's the point of your sayings?

Enough emotions, back to business.