Supermicro H8QGi/6 and H8QGL Next Generation OC BIOS

Status
Not open for further replies.
wow .. that's weird. I have 3 H8QGL -iF+ boards. None of them would take all 3.
Can only use 2 8pin to boot them up and to run (stable with refclock 250+)

Try to remove one to see if it helps :D (if using Y-splitter, then plug it into 8-pin from PSU)
 
Interesting, was there a specific one you left disconnected?

Oh, and I had no problems folding on the stock bios before. These shut downs have only started happening since trying the oc bios and even after flashing back to stock.
 
Last edited:
I use the outer ones .. (right at the edge) ...
Seriously, once the board booted with stock BIOS, i flashed it with tear's OC bios. :D
 
Nice, I'm going to be home soon and can try that out. Hopefully I can keep the system from shutting down during folding. Which opteron are you running a 250 refclock with?
 
might also want to look at dumping that eps splitter and trying out a pci-e to eps adapter
 
6172 set runs at 250
6128 set runs at 259
6174 set runs at 247
they all are using corsair 850tx
 
Wow, I didn't think the 850tx would have been enough for the system. I have one of those in my other system but went up to the 1250 for this build.
 
For 6180SE's it might not be, remember they are 140W TDP. The chips sbinh listed are only 115W TDP.
 
Ah good point. Where can I buy a pci-e to 8pin eps adapter, I had trouble finding them when looking before and only saw 8pin eps to pci-e.
 
I have an overclocked 6180 system running with that same PSU (Seasonic x-1250), and it is plenty of PSU assuming yours is working correctly. I did briefly run one system on a Corsair 750TX, and it worked fine other than running the PSU fan at max rpms. For testing purposes, give your 850 a shot just to rule out a bad PSU.

For the 8 pin MB connectors, I have the two from the PSU plus a molex to 8pin adapter for the third one. Many will say that is a bad idea, but I currently have 74 days uptime at 2813 MHz with no issues.
 
Bad news and I'm not sure how to solve it. I flashed the bios again and now it won't post at all, I tried following the Supermicro steps for recovery, but it's not working. Any suggestions what to try?
 
seven10 said:
Bad news and I'm not sure how to solve it. I flashed the bios again and now it won't post at all, I tried following the Supermicro steps for recovery, but it's not working. Any suggestions what to try?
Click to expand...

I have experienced that a time or two. First time it happened I was sickened and pissed off at the same time. Unplug the PSU/ Let it sit for about 5 to 10 minutes. Clear the CMOS with a paper clip for a good 15 to 20 seconds. Wait another 5 to 10 minutes. Clear the CMOS again for 15 to 20 seconds. Plug the PSU back in and it should power up.

As long as the flash said it was Successful, you should be alright.
 
OCNG doesn't have an option to turn SVM off (IL/PD version will) so you always run with it.
 
Linden said:
Seven10, what's the news? We're with you on this,let us know!
Click to expand...

I bought a Thermaltake PSU and the PSU turned out to be completely fine. The bios had just corrupted probably due to my own re-flashing. Supermicro was able to return that board and send me a replacement extremely fast and I was able to get the system back up and running this week. I have my 6180's stable at 225, but I can't seem to get them above that without failing on work units.
 
Hi tear. I find some motherboards like to increase vcore (0.05V or more) quietly by themself. For example, I manually set the vcore to 1.000V but the value I read with system vcore sensors would be about 1.052V, while the VID value I read is unchanged so it's not affected by the max VID limit.
Is it possible to add such functions to the OCNG BIOS? This would GREATLY improve the OC ability of stock chips.
 
quickz said:
Hi tear. I find some motherboards like to increase vcore (0.05V or more) quietly by themself. For example, I manually set the vcore to 1.000V but the value I read with system vcore sensors would be about 1.052V, while the VID value I read is unchanged so it's not affected by the max VID limit.
Is it possible to add such functions to the OCNG BIOS? This would GREATLY improve the OC ability of stock chips.
Click to expand...

Are you sure its just not the highest common denominator...
Highest vid of the chips installed... it is a range after all.
 
I'm not sure about it yet. I found this when I was comparing the OC behavior of ASUS Z9PE-D16 and SM H8QGL with same chips. I was surprised to see the H8QGL board could let each chip work at the same frequency as Z9PE-D16 with 0.05-0.07V lower vcore setting. I was pleased with it at first, but when I look at the IPMI sensor page I find this might just because the H8QGL board increases the real voltage by itself, the voltage value I read is about 0.05V higher than I set. Would this also be limited by the chip’s highest vcore?
 
quickz, that's GL's VRM setup right there :D GL tends to overshoot voltage targets by quite
a lot when idle.
Some sockets (1, 4) may get a bit above targets even at load [sic!].

VID _logic_ is the same (can't exceed VID voltage beyond limit). It's just GL VRMs that
are being overly "enthusiastic" :) (which is the reason we like GL more than any other board)

What Linux flavor and version are you running?
 
Many thanks for your explanation! I'm using a highly customized RHEL6 distribution with a 3.2.9 kernel.
 
Ok, then I'd recommend that you do this:
1. Power-Off (pull the plug)
2. Disable IPMI (jumper)
3. Power-On
4. modprobe w83795
5. wget http://darkswarm.org/voltcheck.sh
6. chmod +x ./voltcheck.sh
7. ./voltcheck.sh
8. Observe results :)
 
Last edited:
This is a good tool. It works and shows a voltage result of 1.042-1.044V when system is idle while my vcore settting is 0.9875V.
BTW, the script is quite slow. How about changing "sleep 1" to somthing like "sleep 0.1"?

tear said:
Ok, then I'd recommend that you do this:
1. Power-Off (pull the plug)
2. Disable IPMI (jumper)
3. Power-On
4. modprobe w83795
5. wget http://darkswarm.org/voltcheck.sh
6. chmod +x ./voltcheck.sh
7. ./voltcheck.sh
8. Observe results :)
Click to expand...
 
It's been made like that on purpose so the samples used for averaging are spread in time. Feel free to make adjustments though!
 
Last edited:
Hi, I have a question about the memory frequency behavior under OCNG BIOS.
I'm currently using 1.65V XMP 1600 memory sticks. These sticks are unable to apply the "16" flag, right? Then I'm wondering what frequency it would be running under the OCNG BIOS (with “./smocng.sh 200”)?

tear said:
Supported memory configurations:
DDR3-1333 or faster: refclocks 200 through 250 (effective DDR3-1066 through DDR3-1333, respectively)
DDR3-1600: refclocks 251 through 262 (effective DDR3-1339 through DDR3-1397, respectively)
Click to expand...
According to this description, seems the memory frequency would downgrade to 1066Mhz? But under the stock BIOS it’s running at 1333MHz. Did I misunderstand something?
 
quickz said:
Hi, I have a question about the memory frequency behavior under OCNG BIOS.
I'm currently using 1.65V XMP 1600 memory sticks. These sticks are unable to apply the "16" flag, right? Then I'm wondering what frequency it would be running under the OCNG BIOS (with “./smocng.sh 200”)?
Click to expand...

1333.
 
:)
Good find, Mister!

My mistake, the range should start at 201 there, not 200. I'll fix it momentarily. (EDIT: fixed)

In OCNG3, ./smocng.sh 200 yields XMP frequency (if XMP profile is qualified*, otherwise JEDEC),
up to DDR3-1333.

*) qualified means: 1.5V or 1.35V or 1.2V, with one of standard MTB values; DIMMs with non standard
 MTB values (such as Crucial's DDR3-1866 modules) will be supported in OCNG4
 
Last edited:
Supporting DIMMs of higher frequency would be great! It becomes more and more hard to find low latency 1600 DIMMs nowadays.
Hope the OCNG4 BIOS could be released as soon as possible. :D

tear said:
:)
 MTB values (such as Crucial's DDR3-1866 modules) will be supported in OCNG4
Click to expand...
 
Last edited:
( 1333 c7 1.5v ) vs (1600 c8 1.5v) --> which one is better for 4p OCed rig (refclock > 245) ?
 
quickz said:
Supporting DIMMs of higher frequency would be great! It becomes more and more hard to find low latency 1600 DIMMs nowadays.
Hope the OCNG4 BIOS could be released as soon as possible. :D
Click to expand...
It will happen this weekend if nothing interferes...

( 1333 c7 1.5v ) vs (1600 c8 1.5v) --> which one is better for 4p OCed rig (refclock > 245) ?
Click to expand...
For 241-250 it doesn't matter. For 251+ -- 1600.

Price difference is so insignificant I'd opt for DDR3-1600 CL8 or even (future-proof) DDR3-1866 (with OCNG4).
 
Hi tear. Can you say something about the progress on making overclocking bios for Interlagos CPUs?

P.S. I'm running 6176SE at 240 with the 16 setting now (Crucial Ballistix Tactical 1600 CL8). Temps < 49 deg C and TPF around 12 min for 8101 WU.
 
tear said:
It will happen this weekend if nothing interferes...


For 241-250 it doesn't matter. For 251+ -- 1600.

Price difference is so insignificant I'd opt for DDR3-1600 CL8 or even (future-proof) DDR3-1866 (with OCNG4).
Click to expand...

Thanks Tear.
 
DrFredrik said:
Hi tear. Can you say something about the progress on making overclocking bios for Interlagos CPUs?
Click to expand...
ROM has been reworked so it can accommodate new code.
No roadblocks so far. The only issue is lack of time...
Think mid-December or so.


Here's OCNG4 spoiler instead (configuration utility):
Code: 
$ sudo ./ocng-cu
ocng-cu: OCNG configuration utility (NG4)
     
Detected board: 'H8QGL'
     
Reset OCNG configuration to defaults (no/yes) [no]?
Reference clock (200-262) [240]?
Configure advanced options (no/yes) [yes]?
    Perform multiple refclock sets (not recommended) (no/yes) [no]?
    Configure refclock on warm reset only (not recommended) (no/yes) [no]?
    Dynamic HT configuration (not recommended) (no/yes) [no]?
    Prevent use of XMP profile 1 (no/yes) [no]?
    Allow effective memory frequencies above DDR3-1333 (no/yes) [yes]?
    Allow unsafe CAS latencies (not recommended) (no/yes) [no]?
    Force 1.5V DIMM Vdd (DANGEROUS) (no/yes) [yes]?
    Use alternative HT tuning (no/yes) [no]?
    Relax memory timings (no/yes) [no]?
 
tear said:
No roadblocks so far. The only issue is lack of time...
Click to expand...

Check eBay. I thought I saw someone selling some extra for a good price.

Seriously though, yours and everyone else's continuing effect on this is greatly appreciated :)
 
Very nice, I like all of these improvements and look forward to it eagerly.
One question about IPMI: why it cannot work together with the OCNG BIOS on GL boards?

tear said:
Code: 
$Detected board: 'H8QGL'
     
Reset OCNG configuration to defaults (no/yes) [no]?
Reference clock (200-262) [240]?
Configure advanced options (no/yes) [yes]?
    Perform multiple refclock sets (not recommended) (no/yes) [no]?
    Configure refclock on warm reset only (not recommended) (no/yes) [no]?
    Dynamic HT configuration (not recommended) (no/yes) [no]?
    Prevent use of XMP profile 1 (no/yes) [no]?
    Allow effective memory frequencies above DDR3-1333 (no/yes) [yes]?
    Allow unsafe CAS latencies (not recommended) (no/yes) [no]?
    Force 1.5V DIMM Vdd (DANGEROUS) (no/yes) [yes]?
    Use alternative HT tuning (no/yes) [no]?
    Relax memory timings (no/yes) [no]?
Click to expand...
 
I'm pleased to announce general availability of OCNG4 ROM.
See OP for details.

Note that this release still does not support IL/PD chips.
 
Status
Not open for further replies.
Back
Top