A walk to the AMD side

cecht
cecht
Joined: 7 Mar 18
Posts: 1535
Credit: 2909588738
RAC: 2106627

Holmis wrote:... Unwanted

Holmis wrote:
... Unwanted visitors sticking their noses where they don't belong? ...LOLOL!...(not my cat)

Not your cat. Uh-huh, sure.

I actually did consider our two cats when debating whether to leave the case open, but that host is on top of a tall dresser along with my other host, monitor, surge protector, etc, so I think it (and, of course, the kitties) are safe. I took the open-case concept a step further and turned off the case fans, thereby saving and extra 4 W without affecting core temperatures.  That arrangement worked so well that I tried the same for my other host, but that did not work at all; it's a different case with different cards and they ran a lot hotter.  So I restored that host's modesty and let only my new host run in the buff.

Ideas are not fixed, nor should they be; we live in model-dependent reality.

archae86
archae86
Joined: 6 Dec 05
Posts: 3157
Credit: 7226308262
RAC: 1074816

cecht wrote:It can't think

cecht wrote:
It can't think why the bios change wouldn't take hold in Windows, though I've only flipped the switch under Linux.

My original RFX RX 570, for which I've never flipped the mining switch, currently reports 

Bios Information: 113-57085STB1-W90

My second card of that model, which has had inferior performance to the first, but for which I flipped the switch just before last boot reports:

Bios Information: 113-57085SHB2-W90

These differ, which hints that it noticed I flipped the switch (as does the extra initial delay on first boot after the change).

Do you have access to BIOS version information on yours?

 

 

cecht
cecht
Joined: 7 Mar 18
Posts: 1535
Credit: 2909588738
RAC: 2106627

archae86 wrote:cecht wrote:It

archae86 wrote:
cecht wrote:
It can't think why the bios change wouldn't take hold in Windows, though I've only flipped the switch under Linux.

My original RFX RX 570, for which I've never flipped the mining switch, currently reports 

Bios Information: 113-57085STB1-W90

My second card of that model, which has had inferior performance to the first, but for which I flipped the switch just before last boot reports:

Bios Information: 113-57085SHB2-W90

These differ, which hints that it noticed I flipped the switch (as does the extra initial delay on first boot after the change).

Do you have access to BIOS version information on yours?

Good idea. Both my cards, currently running the mining bios, have a vbios of 113-57045EHD1-M90, as reported by amdgpu-utils.  I'm not sure this is the relevant mining/gaming BIOS ID though, because the TechPowerUp GPU Database for that card, https://www.techpowerup.com/vgabios/196582/196582, reports the VBIOS Version as: 015.050.002.001.000000, which is the same general format of bios versions reported by GPU-Z on Windows systems. (I don't know of a GPU-Z equivalent for Linux, so I can't pull up that alternate bios version number for my two cards.)

I don't know what the different bios ID's refer to. I suspect that the one we're seeing is more like a vendor-model ID, in which case it wouldn't change with a flip of the switch. I assume the one used by TechPowerUp is the ID for the gaming bios because the memory timing tables listed under the BIOS Internals on that page don't list the 1850MHz that my 'mining' cards are currently running. Also, my two cards are slightly different, XXX vs Black Edition, with one having a slightly higher "overclock" maximum speed. I would think such a difference would show up in any bios that's running clock, voltages, and timings.

Where did you get your bios IDs from?  If not GPU-Z, then maybe try that utility to compare the alternate bios version IDs with what's listed at TechPowerUp.

Ideas are not fixed, nor should they be; we live in model-dependent reality.

archae86
archae86
Joined: 6 Dec 05
Posts: 3157
Credit: 7226308262
RAC: 1074816

cecht wrote:Where did you get

cecht wrote:
Where did you get your bios IDs from?  If not GPU-Z, then maybe try that utility to compare the alternate bios version IDs with what's listed at TechPowerUp.

The strings I posted as "Bios Information" were provided directly by Windows, with that designation.

I tried looking at both my cards with both GPU-Z and HWInfo, and got these reports:

First card (the better performing one, on which I've never flipped the switch):


Micron memory

Video BIOS Version: 015.050.002.001.000000

ASIC Quality: 73.9%

SMU Firmware Version: 2.23.17


 


Second card (poorer performing in all conditions, switch currently flipped)

015.050.002.001.000000


Hynix memory


Video BIOS Version: 015.050.002.001


ASIC Quality: 73.3%


SMU Firmware Version: 2.23.17


 


I guess sometime soon I'll need to flip the switch back and see if any of these reports change.


But first I should just do a simple full power down and reboot. 


 


Footnote: The site software habit of putting extra linefeeds into my posts drives me nuts.
Holmis
Joined: 4 Jan 05
Posts: 1118
Credit: 1055935564
RAC: 0

archae86 wrote:Footnote: The

archae86 wrote:
Footnote: The site software habit of putting extra linefeeds into my posts drives me nuts.

Try holding down shift while hitting enter for a new line instead of a new paragraph.

Gary Roberts
Gary Roberts
Moderator
Joined: 9 Feb 05
Posts: 5872
Credit: 117696922474
RAC: 35096296

It's really worthwhile

It's really worthwhile reading the BBCode Help drop-down menu item right at the bottom of this page. The coverage of things you can do is quite extensive.  I found the following little gem quite a while ago and it helped enormously with a number of frustrations I was having with formatting.

Quote:
When using the Rich-text editor, loaded by default, ENTER will place a new paragraph and SHIFT-ENTER will place a new line. There are some BBCode tags that will not work with ENTER (new paragraph), e.g., the [list] tag. In general, when entering a line break within a BBCode tag, please use SHIFT-ENTER.

 

Cheers,
Gary.

cecht
cecht
Joined: 7 Mar 18
Posts: 1535
Credit: 2909588738
RAC: 2106627

CEcht wrote:I actually did

CEcht wrote:
I actually did consider our two cats when debating whether to leave the case open, but that host is on top of a tall dresser along with my other host, monitor, surge protector, etc, so I think it (and, of course, the kitties) are safe. I took the open-case concept a step further and turned off the case fans, thereby saving and extra 4 W without affecting core temperatures.

Okay, I'm quoting myself, and for a fairly minor contribution, but this does peripherally concern AMD cards. 

I noticed that the dual fan RX570 cards blow hot air out in pretty much all directions, but more so out the backend of the cards (toward the front of the case). So I flipped the two case fans, reversing the air flow so that it now flows from back-to-front of the case. Then I slightly upped the fan speed profiles in UEFI-bios, replaced the case side panel, et voila!, got decent GPU crunching temperatures. Temps are much better than initially when case fans were working against GPU fans. Added benefit is there's much less GPU heat rising to the CPU. I'm burning a few more Watts running the case fans, but, for now, the sense of achievement seems worth it. Out waste heat, out!

Ideas are not fixed, nor should they be; we live in model-dependent reality.

mikey
mikey
Joined: 22 Jan 05
Posts: 12693
Credit: 1839099411
RAC: 3719

cecht wrote:CEcht wrote:I

cecht wrote:
CEcht wrote:
I actually did consider our two cats when debating whether to leave the case open, but that host is on top of a tall dresser along with my other host, monitor, surge protector, etc, so I think it (and, of course, the kitties) are safe. I took the open-case concept a step further and turned off the case fans, thereby saving and extra 4 W without affecting core temperatures.

Okay, I'm quoting myself, and for a fairly minor contribution, but this does peripherally concern AMD cards. 

I noticed that the dual fan RX570 cards blow hot air out in pretty much all directions, but more so out the backend of the cards (toward the front of the case). So I flipped the two case fans, reversing the air flow so that it now flows from back-to-front of the case. Then I slightly upped the fan speed profiles in UEFI-bios, replaced the case side panel, et voila!, got decent GPU crunching temperatures. Temps are much better than initially when case fans were working against GPU fans. Added benefit is there's much less GPU heat rising to the CPU. I'm burning a few more Watts running the case fans, but, for now, the sense of achievement seems worth it. Out waste heat, out!

Those fans seem counter to normal thinking, I'm glad you figured that out!!

archae86
archae86
Joined: 6 Dec 05
Posts: 3157
Credit: 7226308262
RAC: 1074816

A set of observations on CPU

A set of observations on CPU affinity restrictions:

I recently raised the multiplicity on my second RX 570 from 2X to 3X, and got a nice little productivity improvement, with small enough power consumption increase to make me decide to keep it.  While I was fiddling, I decided to spend a couple of days running restriction (by the CPU affinity controls in Process Lasso) to varying numbers of cores on the 6-core non-hyperthreaded i5-9400F.  I averaged each test condition over several hours.  I continue to run the RX 570 at a -20% power limitation, imposed by MSIAfterburner, and with an Afterburner fan curve which has it reporting 62C GPU temperature most of the time.  This is an XFX brand RX 570 with the BIOS switch on the "mining" position.

The results were simple: restricting to a single core does noticeable harm, but the other five options are surprisingly similar with the (slightly) best result observed with three allowed cores (the same as the number of GPU tasks).  My longstanding observation that restricting the GPU support task to anything less than all available cores always does harm was not borne out.  This does support my long-standing advice that it is better to test than just to invoke "known truths" for these settings.

Cores   Average elapsed time
1       31:14
2       30:18
3       30:12
4       30:14
5       30:16
6       30:17
Peter van Kalleveen
Peter van Kalleveen
Joined: 15 Jan 19
Posts: 45
Credit: 250329645
RAC: 0

archae86 wrote:A set of

archae86 wrote:

A set of observations on CPU affinity restrictions:

I recently raised the multiplicity on my second RX 570 from 2X to 3X, and got a nice little productivity improvement, with small enough power consumption increase to make me decide to keep it.  While I was fiddling, I decided to spend a couple of days running restriction (by the CPU affinity controls in Process Lasso) to varying numbers of cores on the 6-core non-hyperthreaded i5-9400F.  I averaged each test condition over several hours.  I continue to run the RX 570 at a -20% power limitation, imposed by MSIAfterburner, and with an Afterburner fan curve which has it reporting 62C GPU temperature most of the time.  This is an XFX brand RX 570 with the BIOS switch on the "mining" position.

The results were simple: restricting to a single core does noticeable harm, but the other five options are surprisingly similar with the (slightly) best result observed with three allowed cores (the same as the number of GPU tasks).  My longstanding observation that restricting the GPU support task to anything less than all available cores always does harm was not borne out.  This does support my long-standing advice that it is better to test than just to invoke "known truths" for these settings.

Cores   Average elapsed time
1       31:14
2       30:18
3       30:12
4       30:14
5       30:16
6       30:17

How does this respond to the settings of amount of cpu threads you can specify in the boinc App_config? Did you vary this with the affinity or did you use the same ratio while testing?

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.