Bitcoin Forum
June 24, 2021, 04:44:09 PM *
News: Latest Bitcoin Core release: 0.21.1 [Torrent]
 
   Home   Help Search Login Register More  
Pages: « 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 [37] 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 »
  Print  
Author Topic: AMD Mem Tweak XL - Read/modify timings/pp/straps on the fly  (Read 57253 times)
jimmyD30
Jr. Member
*
Offline Offline

Activity: 64
Merit: 1


View Profile
May 21, 2019, 09:29:10 PM

Hello, is anyone having an issue where the last GPU of a multi-GPU rig does not get the mem tweak applied? On 3 out of my 4 rigs (6 Vega64s per rig), the last GPU does not get the tweak applied to it. I知 using CLI (last version) on Win10.
1624553049
Hero Member
*
Offline Offline

Posts: 1624553049

View Profile Personal Message (Offline)

Ignore
1624553049
Reply with quote  #2

1624553049
Report to moderator
1624553049
Hero Member
*
Offline Offline

Posts: 1624553049

View Profile Personal Message (Offline)

Ignore
1624553049
Reply with quote  #2

1624553049
Report to moderator
1624553049
Hero Member
*
Offline Offline

Posts: 1624553049

View Profile Personal Message (Offline)

Ignore
1624553049
Reply with quote  #2

1624553049
Report to moderator
Advertised sites are not endorsed by the Bitcoin Forum. They may be unsafe, untrustworthy, or illegal in your jurisdiction. Advertise here.
joseph32
Member
**
Offline Offline

Activity: 381
Merit: 16


View Profile
May 21, 2019, 09:34:26 PM

vega samsung memory : ETH-CNR-BEAM

WinAMDTweak.exe --RAS 28 --RCDRD 12 --RCDWR 5 --RC 44 --RP 12 --RRDS 3 --RRDL 3 --REF 15600 --RFC 246 --FAW 14


Hey Livada, nice results.
Are these on reference 64's?
What clocks were you running them at per algo?
2.48 on CN/r is nice...

Sorry, but 1 accepted and a 3 minutes run shows nothing. These results with 24 hours would be nice, but I dont bet on it.
SamAlackass
Newbie
*
Offline Offline

Activity: 27
Merit: 1


View Profile
May 21, 2019, 09:52:21 PM

Hello, is anyone having an issue where the last GPU of a multi-GPU rig does not get the mem tweak applied? On 3 out of my 4 rigs (6 Vega64s per rig), the last GPU does not get the tweak applied to it. I知 using CLI (last version) on Win10.

Double-check you enumerate your GPUs correctly. I know it sounds like a silly suggestion, but you know... been there, done that.

If the onboard GPU is anything AMD, it gets listed too (usually GPU0).  
Ravenoss
Newbie
*
Offline Offline

Activity: 3
Merit: 0


View Profile
May 21, 2019, 10:16:55 PM

I have a Win 10 - Vega 56 rig

One Hynix card causes a dead GPU when trying 19.4.2, latest TR and have tried both of the suggested Hynix setttings.  Anyone have one that is working for them?   I've tried:

WinAMDTweak.exe --gpu 03 --RAS 24 --RCDRD 19 --RCDWR 4 --RC 35 --RP 13 --RRDS 4 --RRDL 5 --RFC 148 --REF 15600

and

WinAMDTweak.exe --gpu 03 --RAS 22 --RCDRD 17 --RCDWR 4 --RC 35 --RP 13 --RRDS 4 --RRDL 4 --RFC 148 --REF 15600

Both of these cause a dead gpu/restart
try this-

--rcdrd 19 --rcdwr 4 --rc 35 --rp 14 --rrds 4 --rrdl 5 --rfc 148 --REF 15600
or
--rp 14 --rc 42 --rfc 156 --rrd 3 --rcdrd 18 --rcdwr 3

not best but work

iff any have good setup for vega56/stockbios/hynix pls send ..

I got vega 56 strix hynix and i cant go more than 1800hs.. What is ur best results guys? Any good setup for this bad cards? Sad Huh
pbfarmer
Member
**
Offline Offline

Activity: 340
Merit: 29


View Profile
May 21, 2019, 11:03:36 PM

vega samsung memory : ETH-CNR-BEAM

WinAMDTweak.exe --RAS 28 --RCDRD 12 --RCDWR 5 --RC 44 --RP 12 --RRDS 3 --RRDL 3 --REF 15600 --RFC 246 --FAW 14


Hey Livada, nice results.
Are these on reference 64's?
What clocks were you running them at per algo?
2.48 on CN/r is nice...

Sorry, but 1 accepted and a 3 minutes run shows nothing. These results with 24 hours would be nice, but I dont bet on it.

While it痴 true a 3 min run proves nothing, 2.45+ cnr on a v64 is easily stable at 1375 (measured) / 1107 / 825-850mv using those timings and TRM
jimmyD30
Jr. Member
*
Offline Offline

Activity: 64
Merit: 1


View Profile
May 21, 2019, 11:42:46 PM

Hello, is anyone having an issue where the last GPU of a multi-GPU rig does not get the mem tweak applied? On 3 out of my 4 rigs (6 Vega64s per rig), the last GPU does not get the tweak applied to it. I知 using CLI (last version) on Win10.

Double-check you enumerate your GPUs correctly. I know it sounds like a silly suggestion, but you know... been there, done that.

If the onboard GPU is anything AMD, it gets listed too (usually GPU0).  

I hear ya, made plenty of stupid mistakes myself. But yeah, TRM miner shows the bus number and changes I make to the other GPUs do take effect. I知 wondering if it痴 because I知 using CLI and the - -i parameter and it痴 not reading the last GPU in the parameter setting?
SamAlackass
Newbie
*
Offline Offline

Activity: 27
Merit: 1


View Profile
May 22, 2019, 12:50:05 AM

Hello, is anyone having an issue where the last GPU of a multi-GPU rig does not get the mem tweak applied? On 3 out of my 4 rigs (6 Vega64s per rig), the last GPU does not get the tweak applied to it. I知 using CLI (last version) on Win10.

Double-check you enumerate your GPUs correctly. I know it sounds like a silly suggestion, but you know... been there, done that.

If the onboard GPU is anything AMD, it gets listed too (usually GPU0).  

I hear ya, made plenty of stupid mistakes myself. But yeah, TRM miner shows the bus number and changes I make to the other GPUs do take effect. I知 wondering if it痴 because I知 using CLI and the - -i parameter and it痴 not reading the last GPU in the parameter setting?

You don't happen to start TRM with a -d in config?
jimmyD30
Jr. Member
*
Offline Offline

Activity: 64
Merit: 1


View Profile
May 22, 2019, 02:27:16 AM

Hello, is anyone having an issue where the last GPU of a multi-GPU rig does not get the mem tweak applied? On 3 out of my 4 rigs (6 Vega64s per rig), the last GPU does not get the tweak applied to it. I知 using CLI (last version) on Win10.

Double-check you enumerate your GPUs correctly. I know it sounds like a silly suggestion, but you know... been there, done that.

If the onboard GPU is anything AMD, it gets listed too (usually GPU0).  

I hear ya, made plenty of stupid mistakes myself. But yeah, TRM miner shows the bus number and changes I make to the other GPUs do take effect. I知 wondering if it痴 because I知 using CLI and the - -i parameter and it痴 not reading the last GPU in the parameter setting?

You don't happen to start TRM with a -d in config?

No, but I知 going to do some further tests tomorrow with the latest version and see what happens, maybe double check the settings with the GUI (I知 using CLI right now).
pbfarmer
Member
**
Offline Offline

Activity: 340
Merit: 29


View Profile
May 22, 2019, 03:15:16 AM
Last edit: May 22, 2019, 06:49:32 AM by pbfarmer

Hello, is anyone having an issue where the last GPU of a multi-GPU rig does not get the mem tweak applied? On 3 out of my 4 rigs (6 Vega64s per rig), the last GPU does not get the tweak applied to it. I知 using CLI (last version) on Win10.

Double-check you enumerate your GPUs correctly. I know it sounds like a silly suggestion, but you know... been there, done that.

If the onboard GPU is anything AMD, it gets listed too (usually GPU0).  

I hear ya, made plenty of stupid mistakes myself. But yeah, TRM miner shows the bus number and changes I make to the other GPUs do take effect. I知 wondering if it痴 because I知 using CLI and the - -i parameter and it痴 not reading the last GPU in the parameter setting?

You don't happen to start TRM with a -d in config?

No, but I知 going to do some further tests tomorrow with the latest version and see what happens, maybe double check the settings with the GUI (I知 using CLI right now).

You're using '--i' and not '-i', correct?  The example usage in the help text (amdmemtweak --help) shows the latter, which is not valid.  (EDIT: this was the case in older versions - appears to be fixed in the latest)
livada
Newbie
*
Offline Offline

Activity: 416
Merit: 0


View Profile WWW
May 22, 2019, 06:23:45 AM
Last edit: May 22, 2019, 07:18:14 AM by livada

vega samsung memory : ETH-CNR-BEAM

WinAMDTweak.exe --RAS 28 --RCDRD 12 --RCDWR 5 --RC 44 --RP 12 --RRDS 3 --RRDL 3 --REF 15600 --RFC 246 --FAW 14

http://imgbox.com/oH0Nn1Pa

Hi livada, thanks for all your posts and suggestions, I知 now mining 20-25% more hash at the pool because of you (and of course Eliovp) :-)

But when you run settings tests, you need to run them longer. 1 accepted share does not make a for a very good basis for overall long-term results. Valid results would be based on hours of mining, sometimes GPU crashes with new settings after a while, sometimes many hw errors occur over time, these may not present themselves until many hours of mining.

Cheers!

I test this  1-2 day. This is only SS for send..
With this amd mem tweak setup my 3*vega56 powercolor and 1*vega64LE (all samsung mem) work fine with many algo- I test  CRN-GRAFT- TRTL- MOX-WOW-MSR-LITEv7. any algo have better hr if i change mem tweak setup but with this  setings all algo work fine.

I have a Win 10 - Vega 56 rig

One Hynix card causes a dead GPU when trying 19.4.2, latest TR and have tried both of the suggested Hynix setttings.  Anyone have one that is working for them?   I've tried:

WinAMDTweak.exe --gpu 03 --RAS 24 --RCDRD 19 --RCDWR 4 --RC 35 --RP 13 --RRDS 4 --RRDL 5 --RFC 148 --REF 15600

and

WinAMDTweak.exe --gpu 03 --RAS 22 --RCDRD 17 --RCDWR 4 --RC 35 --RP 13 --RRDS 4 --RRDL 4 --RFC 148 --REF 15600

Both of these cause a dead gpu/restart
try this-

--rcdrd 19 --rcdwr 4 --rc 35 --rp 14 --rrds 4 --rrdl 5 --rfc 148 --REF 15600
or
--rp 14 --rc 42 --rfc 156 --rrd 3 --rcdrd 18 --rcdwr 3

not best but work

iff any have good setup for vega56/stockbios/hynix pls send ..

I got vega 56 strix hynix and i cant go more than 1800hs.. What is ur best results guys? Any good setup for this bad cards? Sad Huh

my 6*veg56 nitro+(hynix) work fine with this setings (+2000hr)- 1 card give 1 error every 1 hour but work fine. itest this with more algo.

WinAMDTweak.exe --rp 14 --rc 35 --rfc 178 --rrds 4 --rrdl 5 --rcdrd 19 --rcdwr 5 --REF 15600

if not work try add --RAS 44(gives stability and lower HR) or chg --rfc 178 to --rfc 188(lower HR)

best card  have gpu= 1440mhz/866mv mem=935mhz/868mv
bad card have   gpu= 1400mhz/870mv mem=895mhz/875mv and give error every 1hour
http://imgbox.com/4Ast1m6p

Ravenoss
Newbie
*
Offline Offline

Activity: 3
Merit: 0


View Profile
May 22, 2019, 07:41:12 AM

vega samsung memory : ETH-CNR-BEAM

WinAMDTweak.exe --RAS 28 --RCDRD 12 --RCDWR 5 --RC 44 --RP 12 --RRDS 3 --RRDL 3 --REF 15600 --RFC 246 --FAW 14

http://imgbox.com/oH0Nn1Pa

Hi livada, thanks for all your posts and suggestions, I知 now mining 20-25% more hash at the pool because of you (and of course Eliovp) :-)

But when you run settings tests, you need to run them longer. 1 accepted share does not make a for a very good basis for overall long-term results. Valid results would be based on hours of mining, sometimes GPU crashes with new settings after a while, sometimes many hw errors occur over time, these may not present themselves until many hours of mining.

Cheers!

I test this  1-2 day. This is only SS for send..
With this amd mem tweak setup my 3*vega56 powercolor and 1*vega64LE (all samsung mem) work fine with many algo- I test  CRN-GRAFT- TRTL- MOX-WOW-MSR-LITEv7. any algo have better hr if i change mem tweak setup but with this  setings all algo work fine.

I have a Win 10 - Vega 56 rig

One Hynix card causes a dead GPU when trying 19.4.2, latest TR and have tried both of the suggested Hynix setttings.  Anyone have one that is working for them?   I've tried:

WinAMDTweak.exe --gpu 03 --RAS 24 --RCDRD 19 --RCDWR 4 --RC 35 --RP 13 --RRDS 4 --RRDL 5 --RFC 148 --REF 15600

and

WinAMDTweak.exe --gpu 03 --RAS 22 --RCDRD 17 --RCDWR 4 --RC 35 --RP 13 --RRDS 4 --RRDL 4 --RFC 148 --REF 15600

Both of these cause a dead gpu/restart
try this-

--rcdrd 19 --rcdwr 4 --rc 35 --rp 14 --rrds 4 --rrdl 5 --rfc 148 --REF 15600
or
--rp 14 --rc 42 --rfc 156 --rrd 3 --rcdrd 18 --rcdwr 3

not best but work

iff any have good setup for vega56/stockbios/hynix pls send ..

I got vega 56 strix hynix and i cant go more than 1800hs.. What is ur best results guys? Any good setup for this bad cards? Sad Huh

my 6*veg56 nitro+(hynix) work fine with this setings (+2000hr)- 1 card give 1 error every 1 hour but work fine. itest this with more algo.

WinAMDTweak.exe --rp 14 --rc 35 --rfc 178 --rrds 4 --rrdl 5 --rcdrd 19 --rcdwr 5 --REF 15600

if not work try add --RAS 44(gives stability and lower HR) or chg --rfc 178 to --rfc 188(lower HR)

best card  have gpu= 1440mhz/866mv mem=935mhz/868mv
bad card have   gpu= 1400mhz/870mv mem=895mhz/875mv and give error every 1hour
http://imgbox.com/4Ast1m6p


What miner u use? I use last release of Srbminer and results very bad  Cry
livada
Newbie
*
Offline Offline

Activity: 416
Merit: 0


View Profile WWW
May 22, 2019, 08:03:21 AM

What miner u use? I use last release of Srbminer and results very bad  Cry

TRM for all suport(v8), phoenix for eth, gminer for beam and srb for other  alfgo like litev7-mox-wow
Iamtutut
Full Member
***
Offline Offline

Activity: 1106
Merit: 128


View Profile
May 22, 2019, 10:50:25 AM

New Release

  • Added FIJI support
  • Added Refresh Rate timing (RX boost  Wink)
  • Overdrive fixes
  • Added memory manufacturer detection for G5, HBM & HBM2
  • .....

Link: Here


Will upload CLI versions asap as well!


I noticed some people have figured out that there is still room to improve on RX 4**/5** cards.
Just not many people saying how..

Well, open up the tool, and raise Refresh Rate (REF) i created a separate textbox for it to make it easy.
Raise it from 5 to 7/8 (just an example) and you'll see improvement already.

Slightly dropping RFC will increase your hashrate as well.



Greetings

It's slightly faster, thanks a lot !
I applied my usual settings for RX574.

tRFC:
- Samsung memory (tCL is 22): 117
- Hynix memory (tCL is 19): 124
- Elpida memory (tCL is 21): 94

Right now refresh at 9 for all GPUs. I'll see if it's stable.

These settings rae very stable to me, rehresf at 8 for Samsung memory, 9 for Hynix and Elpida.

Eliovp, do you accept tips in TRTL (turtlecoin) ?
jimmyD30
Jr. Member
*
Offline Offline

Activity: 64
Merit: 1


View Profile
May 22, 2019, 01:01:25 PM

Hello, is anyone having an issue where the last GPU of a multi-GPU rig does not get the mem tweak applied? On 3 out of my 4 rigs (6 Vega64s per rig), the last GPU does not get the tweak applied to it. I知 using CLI (last version) on Win10.

Double-check you enumerate your GPUs correctly. I know it sounds like a silly suggestion, but you know... been there, done that.

If the onboard GPU is anything AMD, it gets listed too (usually GPU0).  

I hear ya, made plenty of stupid mistakes myself. But yeah, TRM miner shows the bus number and changes I make to the other GPUs do take effect. I知 wondering if it痴 because I知 using CLI and the - -i parameter and it痴 not reading the last GPU in the parameter setting?

You don't happen to start TRM with a -d in config?

No, but I知 going to do some further tests tomorrow with the latest version and see what happens, maybe double check the settings with the GUI (I知 using CLI right now).

You're using '--i' and not '-i', correct?  The example usage in the help text (amdmemtweak --help) shows the latter, which is not valid.  (EDIT: this was the case in older versions - appears to be fixed in the latest)


Thank you for the suggestion, yes I知 using the double dash.
jimmyD30
Jr. Member
*
Offline Offline

Activity: 64
Merit: 1


View Profile
May 22, 2019, 01:06:28 PM

I test this  1-2 day. This is only SS for send..
With this amd mem tweak setup my 3*vega56 powercolor and 1*vega64LE (all samsung mem) work fine with many algo- I test  CRN-GRAFT- TRTL- MOX-WOW-MSR-LITEv7. any algo have better hr if i change mem tweak setup but with this  setings all algo work fine.


Thank you for the clarification :-)
Ravenoss
Newbie
*
Offline Offline

Activity: 3
Merit: 0


View Profile
May 22, 2019, 01:59:29 PM

What miner u use? I use last release of Srbminer and results very bad  Cry

TRM for all suport(v8), phoenix for eth, gminer for beam and srb for other  alfgo like litev7-mox-wow

Thank you very much, i got 2000hs for card, very good miner and timings Smiley
zeo229
Newbie
*
Offline Offline

Activity: 5
Merit: 0


View Profile
May 22, 2019, 02:16:11 PM

I can't run the program (gui). Get an excetpion:
StartService failed: The service cannot be started,
either because it is disabled or because it has no enabled devices associated with it
at WinFormsGUI.IOMap.Loaddriver()
at WinFormsGUI.Program.Main()



What could be the issue?
rednoW
Legendary
*
Offline Offline

Activity: 1514
Merit: 1003


View Profile
May 22, 2019, 02:31:29 PM

Is it only for me that loading profile with latest V5 results in fan speeds blank table at the right bottom ?
Polaris cards. V4 was and is OK.
pbfarmer
Member
**
Offline Offline

Activity: 340
Merit: 29


View Profile
May 22, 2019, 05:15:10 PM

New Release

  • Added FIJI support
  • Added Refresh Rate timing (RX boost  Wink)
  • Overdrive fixes
  • Added memory manufacturer detection for G5, HBM & HBM2
  • .....

Link: Here


Will upload CLI versions asap as well!


I noticed some people have figured out that there is still room to improve on RX 4**/5** cards.
Just not many people saying how..

Well, open up the tool, and raise Refresh Rate (REF) i created a separate textbox for it to make it easy.
Raise it from 5 to 7/8 (just an example) and you'll see improvement already.

Slightly dropping RFC will increase your hashrate as well.



Greetings

It's slightly faster, thanks a lot !
I applied my usual settings for RX574.

tRFC:
- Samsung memory (tCL is 22): 117
- Hynix memory (tCL is 19): 124
- Elpida memory (tCL is 21): 94

Right now refresh at 9 for all GPUs. I'll see if it's stable.

These settings rae very stable to me, rehresf at 8 for Samsung memory, 9 for Hynix and Elpida.

Eliovp, do you accept tips in TRTL (turtlecoin) ?

I知 currently benching  REF 10 across all my 580s (samsung, micron, and hynix) - 3 errors out of 1496 results total (.2%) so far.  They are actually from 2 cards specifically - a Samsung (2/192) and a Hynix (1/171).

Gain looks to be ~0.7-0.8Mh/s (ethash) per card, or about 2-2.5%
tale01
Newbie
*
Offline Offline

Activity: 114
Merit: 0


View Profile
May 22, 2019, 05:24:31 PM

What config for Fiji ?  Huh

I follow picture on github, it cannot works.
Pages: « 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 [37] 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 »
  Print  
 
Jump to:  

Powered by MySQL Powered by PHP Powered by SMF 1.1.19 | SMF © 2006-2009, Simple Machines Valid XHTML 1.0! Valid CSS!