Overclock.net › Forums › Intel › Intel Motherboards › Review of GIGABYTE's 2nd Gen. X79 Workstation MB: X79S-UP5
New Posts  All Forums:Forum Nav:

Review of GIGABYTE's 2nd Gen. X79 Workstation MB: X79S-UP5 - Page 15

post #141 of 367
Quote:
What voltages are you attempting?
Quote:
Are you using the XMP profile?

What BIOS version?
@Blameless : At first, voltage was set to AUTO. I was seeing around 1.32v with the multiplier set to 45x. Even when I decided to change the voltage to 1.28v with a multiplier of 45x (everything else on AUTO), it still behaves the same.

Edit: Its bios F4s
Quote:
As for Everest (AIDA64) benchmark, yes, it does take the same amount of time for me on this board. This is actually "by design" believe it or not, at least according to AIDA64. Apparently due to the complexity of the IMC and the pathways it can take several minutes for the benchmark to calculate latencies accurately.
@dejanh: ok, thanks for the info. I though it was just me. I'll give this a go again with a different board/maker.

I wonder if a Xeon would behave the same if you raise the base clock? Gonna try this...
Edited by maxxx.ph - 12/5/12 at 8:09pm
RX-0
(14 items)
 
  
Reply
RX-0
(14 items)
 
  
Reply
post #142 of 367
Quote:
Originally Posted by Blameless View Post

Yeah, I did notice this.
This proves that it's not just a current limit.
EDIT: The 165A current limit on SB-E parts corresponds almost exactly to the 193w peak I was seeing at the 1.2-1.25v I was attempting to OC with. So, it appears that there is both a current and total wattage limit, just as the BIOS settings would suggest, but no way to actually increase them.

Intel has specifications for max current (165A for the cores of hex core parts) as well, so it seems odd to me that you would exceed that by nearly three fold while sticking closely to the max voltage rating.
I think a more reasonable limit would be based off a minimum of double the current specification and 1.6v minimum, to account for overclocking with high end cooling. So, no hard limit until at least 330A and 528w. And obviously these limits should be reflected in the BIOS settings.
I do understand why Gigabyte set the limit lower. The VRM already gets rather warm at around 120A, and 300A+ would not be sustainable without active cooling. Still, there is absolutely no reason for a cap as low as it is, on a board that advertises OCing ability on the front of the box. There is also no reason for a 1200w limit to be an option in the BIOS when the real limit is less than one sixth that.
I see what you are saying though I'm still convinced that the limits that I proposed are quite reasonable. I've measured the VRM temperatures and they come in at about 40C - 45C right now when the chip is fully stressed at my maximum possible overclock of 4.0GHz (about 145A draw). This is nothing to fret at as these suckers will run up to 135C. Keep in mind that the temperature should not scale linearly (should be better than linear) as the amperage increases. You certainly do not see a 20% increase in VRM temperatures by loading the chip with 20% more amperage right now. The temperature difference is only a few degrees C. The numbers that I proposed are an extrapolation based on the straight math and datasheet specifications. They are perfectly doable. Even if you need some airflow over the VRM it's not the end of the world. Moreover, Gigabyte could implement not just a current threshold but a temperature threshold for the VRM. If they hit 110C start to throttle the chip to keep VRM temperatures within 110C. Setting the current and power limits to what I have suggested would give us plenty of OC headroom. The board would not be winning any OC competitions but it would be an excellent choice for your average enthusiast, especially on water.

So now we have concluded that we have both a current and a power limit in place. This makes a lot of sense, and it also is in line with my assumption that the two settings in the UEFI BIOS for current and power do not work. I am actually questioning whether the other settings related to the power plane work correctly either since I have not been able to see any different when any of them are applied. However, one thing at a time.

On a funny note, what made me think that the current and power settings have no effect is actually the fact that they maxed out at 300A and 1200W. 300A does not divide by 8, so that made no sense. 1200W certainly is not any feasible multiplier of the maximum amperage (you would need 4V at 300A to get 1200W) so that number made no sense either. Basically, right away I noticed that the two settings did not seem to correlate to anything, or to each other which seemed rather odd. Anyway, it's pretty funny smile.gif
Quote:
Originally Posted by maxxx.ph View Post

@Blameless : At first, voltage was set to AUTO. I was seeing around 1.32v with the multiplier set to 45x. Even when I decided to change the voltage to 1.28v with a multiplier of 45x (everything else on AUTO), it still behaves the same.
Edit: Its bios F4s
@dejanh: ok, thanks for the info. I though it was just me. I'll give this a go again with a different board/maker.
I wonder if a Xeon would behave the same if you raise the base clock? Gonna try this...
I'm curious what the Xeon does. I am not sure though how much you will get out of that chip by raising the BCLK. Also, is it an 8C/16T chip or just a plain old 4C/8T? You will definitely not get any throttling with a 4C/8T chip.
Edited by dejanh - 12/5/12 at 8:50pm
Etch-a-Sketch
(16 items)
 
  
CPUMotherboardGraphicsRAM
Intel Core i7 3930K Gigabyte GA-X79S-UP5 2x MSI R5850 OC 1.350V/1060/1260 in CrossFireX 64GB Corsair Dominator 1866MHz, 9-10-9-27-180-1T 
Hard DriveCoolingMonitorKeyboard
2x Corsair Force 120GB GT SSD in RAID-0, 2x WD ... EK Supreme HF (EN), HW Labs BI GTX480, HW Labs ... Samsung S27A850 Logitech G19 Gaming Keyboard 
PowerCaseMouseAudio
Corsair Professional Series Gold CMPSU-1200AX Antec 1200 Gamer Case, Custom Full-window Side ... Logitech G9 Laser Mouse Auzentech X-Fi Forte 
Other
Bigfoot Networks Killer Xeno Pro 
  hide details  
Reply
Etch-a-Sketch
(16 items)
 
  
CPUMotherboardGraphicsRAM
Intel Core i7 3930K Gigabyte GA-X79S-UP5 2x MSI R5850 OC 1.350V/1060/1260 in CrossFireX 64GB Corsair Dominator 1866MHz, 9-10-9-27-180-1T 
Hard DriveCoolingMonitorKeyboard
2x Corsair Force 120GB GT SSD in RAID-0, 2x WD ... EK Supreme HF (EN), HW Labs BI GTX480, HW Labs ... Samsung S27A850 Logitech G19 Gaming Keyboard 
PowerCaseMouseAudio
Corsair Professional Series Gold CMPSU-1200AX Antec 1200 Gamer Case, Custom Full-window Side ... Logitech G9 Laser Mouse Auzentech X-Fi Forte 
Other
Bigfoot Networks Killer Xeno Pro 
  hide details  
Reply
post #143 of 367
Quote:
I'm curious what the Xeon does. I am not sure though how much you will get out of that chip by raising the BCLK. Also, is it an 8C/16T chip or just a plain old 4C/8T? You will definitely not get any throttling with a 4C/8T chip.
It's the 8c/16t (E5-2690). I forgot the model of the other pair. I'll check and try when I get back home.
RX-0
(14 items)
 
  
Reply
RX-0
(14 items)
 
  
Reply
post #144 of 367
Guys

got everything today only but tried a quick setup and some of the test. First off I am no overclocker so not sure if I set everything correctly. I tried to follow some of your instructions as close as I could. I gotta get some sleep before work but will test some more later today - please see the screenshot and let me know if anything looks out of place. To me at looked like no throttling was occurring under load since the multiplier never dropped below 44 while I was running the Linx test. Something does not look right with the time and GFlops though if I compare to your screenshots.



Sorry for the limited information at this time but let me know exactly what you need to see.

Thanks
The Throttler
(14 items)
 
  
CPUMotherboardGraphicsRAM
i7 3930K Gigabyte X79S-UP5-WIFI XFX 9600 GT Corsair Vengeance LP CML16GX3M2A1600C10 x 2 
Hard DriveOptical DriveCoolingCooling
OCZ Agility 4 256GB SH-224BB Corsair Hydro Series H100 Extreme Performance L... Corsair SP120 Twin Pack 
OSMonitorMonitorPower
Windows 8 Dell S2309W Dell E228WFP OCZ ZX1250W 
CaseMouse
Corsair 600TM Alienware Tactx Mouse 
  hide details  
Reply
The Throttler
(14 items)
 
  
CPUMotherboardGraphicsRAM
i7 3930K Gigabyte X79S-UP5-WIFI XFX 9600 GT Corsair Vengeance LP CML16GX3M2A1600C10 x 2 
Hard DriveOptical DriveCoolingCooling
OCZ Agility 4 256GB SH-224BB Corsair Hydro Series H100 Extreme Performance L... Corsair SP120 Twin Pack 
OSMonitorMonitorPower
Windows 8 Dell S2309W Dell E228WFP OCZ ZX1250W 
CaseMouse
Corsair 600TM Alienware Tactx Mouse 
  hide details  
Reply
post #145 of 367
Quote:
Originally Posted by cobretti View Post

Guys
got everything today only but tried a quick setup and some of the test. First off I am no overclocker so not sure if I set everything correctly. I tried to follow some of your instructions as close as I could. I gotta get some sleep before work but will test some more later today - please see the screenshot and let me know if anything looks out of place. To me at looked like no throttling was occurring under load since the multiplier never dropped below 44 while I was running the Linx test. Something does not look right with the time and GFlops though if I compare to your screenshots.

Sorry for the limited information at this time but let me know exactly what you need to see.
Thanks
Hey there,

Thanks for doing the testing. You have the right idea, but the wrong version of LinX. Get the AVX compatible version from here http://www.mediafire.com/?7w8xpkd233i85f8

When running LinX set the problem size to 25,000 and make sure that you watch the CPU-Z clocks carefully if you want to see the throttling. I can already tell you your chip will also throttle since it's already hitting 175W and you aren't even using AVX yet.

Looking forward to new results smile.gif
Etch-a-Sketch
(16 items)
 
  
CPUMotherboardGraphicsRAM
Intel Core i7 3930K Gigabyte GA-X79S-UP5 2x MSI R5850 OC 1.350V/1060/1260 in CrossFireX 64GB Corsair Dominator 1866MHz, 9-10-9-27-180-1T 
Hard DriveCoolingMonitorKeyboard
2x Corsair Force 120GB GT SSD in RAID-0, 2x WD ... EK Supreme HF (EN), HW Labs BI GTX480, HW Labs ... Samsung S27A850 Logitech G19 Gaming Keyboard 
PowerCaseMouseAudio
Corsair Professional Series Gold CMPSU-1200AX Antec 1200 Gamer Case, Custom Full-window Side ... Logitech G9 Laser Mouse Auzentech X-Fi Forte 
Other
Bigfoot Networks Killer Xeno Pro 
  hide details  
Reply
Etch-a-Sketch
(16 items)
 
  
CPUMotherboardGraphicsRAM
Intel Core i7 3930K Gigabyte GA-X79S-UP5 2x MSI R5850 OC 1.350V/1060/1260 in CrossFireX 64GB Corsair Dominator 1866MHz, 9-10-9-27-180-1T 
Hard DriveCoolingMonitorKeyboard
2x Corsair Force 120GB GT SSD in RAID-0, 2x WD ... EK Supreme HF (EN), HW Labs BI GTX480, HW Labs ... Samsung S27A850 Logitech G19 Gaming Keyboard 
PowerCaseMouseAudio
Corsair Professional Series Gold CMPSU-1200AX Antec 1200 Gamer Case, Custom Full-window Side ... Logitech G9 Laser Mouse Auzentech X-Fi Forte 
Other
Bigfoot Networks Killer Xeno Pro 
  hide details  
Reply
post #146 of 367
Quote:
Originally Posted by dejanh View Post

Hey there,
Thanks for doing the testing. You have the right idea, but the wrong version of LinX. Get the AVX compatible version from here http://www.mediafire.com/?7w8xpkd233i85f8
When running LinX set the problem size to 25,000 and make sure that you watch the CPU-Z clocks carefully if you want to see the throttling. I can already tell you your chip will also throttle since it's already hitting 175W and you aren't even using AVX yet.
Looking forward to new results smile.gif
You are welcome and you are spot on - immediately see the throttling now. It did not even complete first run with LinX and after about 40 secs in started throttling and with anything 180W+

So no GB golden chip for me either frown.gif - hope they get someone who can understand your test results and just correct it.
The Throttler
(14 items)
 
  
CPUMotherboardGraphicsRAM
i7 3930K Gigabyte X79S-UP5-WIFI XFX 9600 GT Corsair Vengeance LP CML16GX3M2A1600C10 x 2 
Hard DriveOptical DriveCoolingCooling
OCZ Agility 4 256GB SH-224BB Corsair Hydro Series H100 Extreme Performance L... Corsair SP120 Twin Pack 
OSMonitorMonitorPower
Windows 8 Dell S2309W Dell E228WFP OCZ ZX1250W 
CaseMouse
Corsair 600TM Alienware Tactx Mouse 
  hide details  
Reply
The Throttler
(14 items)
 
  
CPUMotherboardGraphicsRAM
i7 3930K Gigabyte X79S-UP5-WIFI XFX 9600 GT Corsair Vengeance LP CML16GX3M2A1600C10 x 2 
Hard DriveOptical DriveCoolingCooling
OCZ Agility 4 256GB SH-224BB Corsair Hydro Series H100 Extreme Performance L... Corsair SP120 Twin Pack 
OSMonitorMonitorPower
Windows 8 Dell S2309W Dell E228WFP OCZ ZX1250W 
CaseMouse
Corsair 600TM Alienware Tactx Mouse 
  hide details  
Reply
post #147 of 367
Quote:
Originally Posted by dejanh View Post

I see what you are saying though I'm still convinced that the limits that I proposed are quite reasonable. I've measured the VRM temperatures and they come in at about 40C - 45C right now when the chip is fully stressed at my maximum possible overclock of 4.0GHz (about 145A draw). This is nothing to fret at as these suckers will run up to 135C. Keep in mind that the temperature should not scale linearly (should be better than linear) as the amperage increases. You certainly do not see a 20% increase in VRM temperatures by loading the chip with 20% more amperage right now. The temperature difference is only a few degrees C. The numbers that I proposed are an extrapolation based on the straight math and datasheet specifications. They are perfectly doable. Even if you need some airflow over the VRM it's not the end of the world. Moreover, Gigabyte could implement not just a current threshold but a temperature threshold for the VRM. If they hit 110C start to throttle the chip to keep VRM temperatures within 110C. Setting the current and power limits to what I have suggested would give us plenty of OC headroom. The board would not be winning any OC competitions but it would be an excellent choice for your average enthusiast, especially on water.

Actual temp of the VRMs are surely higher than either side of the board, and I've been able to hit 60C (my ambients are in the 30C range) on the backside of the PCB behind the VRMs. Efficency of the VRM also declines with temperature, current (past ~20A per phase), and voltage (past 1.2v output). With each phase potentially needing to disipate 10+ watts of heat, the default heatsink would certainly be overwhelmed at max current draw, and probably quite a bit below.

Anyway, regardless of the specifics, I certainly agree that the current limit in place now is far too low, and that it could be raised to an arbitrarily high level if VRM temps where controled.

It's also interesting to note that the default value of 135A corresponds to Intel's Thermal Design Current, while 124% of that value (which is the max OCP setting in the BIOS) corresponds very well with the apparent 165A limit we are observing, and the 165A maximum processor current in Intel's white paper.
Quote:
Originally Posted by dejanh View Post

So now we have concluded that we have both a current and a power limit in place. This makes a lot of sense, and it also is in line with my assumption that the two settings in the UEFI BIOS for current and power do not work.

I concur 100%.
Quote:
Originally Posted by dejanh View Post

On a funny note, what made me think that the current and power settings have no effect is actually the fact that they maxed out at 300A and 1200W. 300A does not divide by 8, so that made no sense. 1200W certainly is not any feasible multiplier of the maximum amperage (you would need 4V at 300A to get 1200W) so that number made no sense either. Basically, right away I noticed that the two settings did not seem to correlate to anything, or to each other which seemed rather odd.

300 / 8 = 27.5 and there is nothing inherently wrong with this number; amperage does not need to be a whole number.

The watt limit doesn't seem to correlate to anything, this is true, but if they are being measured independently, it doesn't necessarily have to. 1200w is a stupid limit to set, but that doesn't mean it couldn't actually be the limit if they wanted it there.

Anyway, It's pretty clear these settings don't work.
Quote:
Originally Posted by dejanh View Post

I'm curious what the Xeon does. I am not sure though how much you will get out of that chip by raising the BCLK. Also, is it an 8C/16T chip or just a plain old 4C/8T? You will definitely not get any throttling with a 4C/8T chip.

LGA 2011 Xeons are strap locked as well as multiplier locked, so any significant OCing is out of the question anyway.

On a side note, this board does really badly with the 1.25x strap in my experience. No BIOS past F4h would allow me to even boot at 1.25x, and even the earlier revisions needed me to back of to ~24x CPU multiplier. I was also required to use AUTO voltage settings, which overvolted VTT and VCCSA substantially, potentially dangerously for C2 parts.
Quote:
Originally Posted by cobretti View Post

immediately see the throttling now. It did not even complete first run with LinX and after about 40 secs in started throttling and with anything 180W+
So no GB golden chip for me either frown.gif - hope they get someone who can understand your test results and just correct it.

I find it hilarious, in a so-angry-I-could-kill kinda way, that pretty much every single end user who has tested this board with an AVX stress test is seeing throttling, yet none of the nearly dozen reviews I've read of the board on professional review sites have mentioned anything of the sort, even at OCs I cannot even boot at, and that Gigabyte is still barely acknowledging the existence of the issue.
Edited by Blameless - 12/6/12 at 7:30am
Primary
(15 items)
 
Secondary
(13 items)
 
In progress
(10 items)
 
CPUMotherboardGraphicsRAM
5820K @ 4.2/3.5GHz core/uncore, 1.175/1.15v Gigabyte X99 SOC Champion (F22n) Gigabyte AORUS GTX 1080 Ti (F3P) @ 2025/1485, 1... 4x4GiB Crucial @ 2667, 12-12-12-28-T1, 1.34v 
Hard DriveHard DriveHard DriveCooling
Plextor M6e 128GB (fw 1.06) M.2 (PCI-E 2.0 2x) 2x Crucial M4 256GB 4x WD Scorpio Black 500GB Noctua NH-D15 
OSMonitorKeyboardPower
Windows 7 Professional x64 SP1 BenQ BL3200PT Filco Majestouch Tenkeyless (MX Brown) Corsair RM1000x 
CaseMouseAudio
Fractal Design Define R4 Logitech G402 Realtek ALC1150 + M-Audio AV40 
CPUMotherboardGraphicsRAM
X5670 @ 4.4/3.2GHz core/uncore, 1.36 vcore, 1.2... Gigabyte X58A-UD5 r2.0 w/FF3mod10 BIOS Sapphire Fury Nitro OC+ @ 1053/500, 1.225vGPU/1... 2x Samsung MV-3V4G3D/US @ 2000, 10-11-11-30-T1,... 
RAMHard DriveHard DriveHard Drive
1x Crucial BLT4G3D1608ET3LX0 @ 2000, 10-11-11-3... OCZ (Toshiba) Trion 150 120GB Hyundai Sapphire 120GB 3x Hitachi Deskstar 7k1000.C 1TB 
CoolingOSPowerCase
Noctua NH-D14 Windows 7 Pro x64 SP1 Antec TP-750 Fractal Design R5 
Audio
ASUS Xonar DS 
CPUMotherboardGraphicsRAM
i7-6800K @ 4.3/3.5GHz core/uncore, 1.36/1.2v ASRock X99 OC Formula (P3.10) GTX 780 (temporary) 4x4GiB Crucial DDR4-2400 @ 11-13-12-28-T2, 1.33v 
Hard DriveHard DriveCoolingOS
Intel 600p 256GB NVMe 2x HGST Travelstar 7k1000 1TB Corsair H55 (temporary) Windows Server 2016 Datacenter 
PowerCase
Seasonic SS-860XP2 Corsair Carbide Air 540 
  hide details  
Reply
Primary
(15 items)
 
Secondary
(13 items)
 
In progress
(10 items)
 
CPUMotherboardGraphicsRAM
5820K @ 4.2/3.5GHz core/uncore, 1.175/1.15v Gigabyte X99 SOC Champion (F22n) Gigabyte AORUS GTX 1080 Ti (F3P) @ 2025/1485, 1... 4x4GiB Crucial @ 2667, 12-12-12-28-T1, 1.34v 
Hard DriveHard DriveHard DriveCooling
Plextor M6e 128GB (fw 1.06) M.2 (PCI-E 2.0 2x) 2x Crucial M4 256GB 4x WD Scorpio Black 500GB Noctua NH-D15 
OSMonitorKeyboardPower
Windows 7 Professional x64 SP1 BenQ BL3200PT Filco Majestouch Tenkeyless (MX Brown) Corsair RM1000x 
CaseMouseAudio
Fractal Design Define R4 Logitech G402 Realtek ALC1150 + M-Audio AV40 
CPUMotherboardGraphicsRAM
X5670 @ 4.4/3.2GHz core/uncore, 1.36 vcore, 1.2... Gigabyte X58A-UD5 r2.0 w/FF3mod10 BIOS Sapphire Fury Nitro OC+ @ 1053/500, 1.225vGPU/1... 2x Samsung MV-3V4G3D/US @ 2000, 10-11-11-30-T1,... 
RAMHard DriveHard DriveHard Drive
1x Crucial BLT4G3D1608ET3LX0 @ 2000, 10-11-11-3... OCZ (Toshiba) Trion 150 120GB Hyundai Sapphire 120GB 3x Hitachi Deskstar 7k1000.C 1TB 
CoolingOSPowerCase
Noctua NH-D14 Windows 7 Pro x64 SP1 Antec TP-750 Fractal Design R5 
Audio
ASUS Xonar DS 
CPUMotherboardGraphicsRAM
i7-6800K @ 4.3/3.5GHz core/uncore, 1.36/1.2v ASRock X99 OC Formula (P3.10) GTX 780 (temporary) 4x4GiB Crucial DDR4-2400 @ 11-13-12-28-T2, 1.33v 
Hard DriveHard DriveCoolingOS
Intel 600p 256GB NVMe 2x HGST Travelstar 7k1000 1TB Corsair H55 (temporary) Windows Server 2016 Datacenter 
PowerCase
Seasonic SS-860XP2 Corsair Carbide Air 540 
  hide details  
Reply
post #148 of 367
I hope its not another case of the GA-X79-UD7. frown.gif

Anyone notice that the multiplier does not stick to what you set it when the cpu gets into full load? Even with all those EIST, etc set to disabled in the BIOS/UEFI.
Is that normal? It seems to be the case as my other Gigabyte board behaves the same way (UD7). Anyone with any other brand can test?

The Xeon E5 2620 is behaving the same. I've set the multi to 24x but when Linx or Prime95 runs, it goes down to 23x. This one sticks to 23x though, does not go lower. Maybe because it has not reach its ceiling. I'll test with the higher TDP Xeon when I get the chance.


Edited by maxxx.ph - 12/6/12 at 8:45am
RX-0
(14 items)
 
  
Reply
RX-0
(14 items)
 
  
Reply
post #149 of 367
Quote:
Originally Posted by Blameless View Post

Actual temp of the VRMs are surely higher than either side of the board, and I've been able to hit 60C (my ambients are in the 30C range) on the backside of the PCB behind the VRMs. Efficency of the VRM also declines with temperature, current (past ~20A per phase), and voltage (past 1.2v output). With each phase potentially needing to disipate 10+ watts of heat, the default heatsink would certainly be overwhelmed at max current draw, and probably quite a bit below.
Anyway, regardless of the specifics, I certainly agree that the current limit in place now is far too low, and that it could be raised to an arbitrarily high level if VRM temps where controled.
It's also interesting to note that the default value of 135A corresponds to Intel's Thermal Design Current, while 124% of that value (which is the max OCP setting in the BIOS) corresponds very well with the apparent 165A limit we are observing, and the 165A maximum processor current in Intel's white paper.
I agree with what you are saying, though my observations on temperature scaling measured with the probe shoved between two VRMs on the backside of the board underneath the heatsink showed only a nominal increase in temperature. The efficiency drop of this VRM at the amperage increases is also very nominal. At 40A the VRM is 92% efficient, and at 60A it is still 88% efficient. That is very efficient meaning that there should be minimal loss to wasted energy. Couple this with the fact that the datasheet actually specifies maximum operating temperature at 150C (not 135C that we thought), and I really would say that there is nothing to worry about. Even if the VRM operated at 100C daily it's not a big deal. It is still only 67% of the maximum temperature. In my case I always have some air blowing over the VRM as well so I am really not too concerned. We are in agreement though, and that's the most important thing smile.gif
Quote:
Originally Posted by Blameless View Post

300 / 8 = 27.5 and there is nothing inherently wrong with this number; amperage does not need to be a whole number.
The watt limit doesn't seem to correlate to anything, this is true, but if they are being measured independently, it doesn't necessarily have to. 1200w is a stupid limit to set, but that doesn't mean it couldn't actually be the limit if they wanted it there.
Anyway, It's pretty clear these settings don't work.
It's not that there is something "wrong" with the amperage being set at 300A in that sense, it's just not very common for somebody to configure the setting at a value that is not a multiple of the whole. Trust me on this, speaking from experience in dealing with people it's a tell-tale sign of somebody setting an arbitrary number without actually thinking about what it implies. Likewise, 1200W is not just a stupid limit to set, it is an impossible limit to set. The output voltage range for the IR3550 is 0.25V to 3.3V, which means that (with the current limit of 300A and to achieve 1200W of power) the voltage would have to exceed the maximum output voltage of the IR3550 by approximately 22%. Both numbers were set without any meaning or through behind it. Again though, the settings don't work and at the end of the day that's what matters to know. I guess for me it is just a sign of major sloppiness on Gigabyte firmware engineering team. These types of things should never happen and should never make it through the QA. Furthermore, the fact that these settings do not work at all proves in itself that nobody actually bothered to test the features on the board. Keep in mind that these are not the only non-working features either. I am still guaranteeing that the lack of tRC setting and the artificially imposed limit of "32" for manual setting is nothing but a blatant oversight in the design of the firmware.
Quote:
Originally Posted by Blameless View Post

I find it hilarious, in a so-angry-I-could-kill kinda way, that pretty much every single end user who has tested this board with an AVX stress test is seeing throttling, yet none of the nearly dozen reviews I've read of the board on professional review sites have mentioned anything of the sort, even at OCs I cannot even boot at, and that Gigabyte is still barely acknowledging the existence of the issue.
We'll see how long they can do this denial dance. Every user of a C2 chip on this board will experience throttling. There are no exceptions. This effectively means that every retail buyer of an unlocked SB-E chip and this board will experience this problem. That is 100% of users. Not 1%, not 5%, not 50%, a full 100% of users experiencing the same problem. As for reviewers, shows how thorough and objective they are. Until reviewers start buying their own gear and reviewing it they are nothing but a bunch of marketing tools for manufacturers/resellers. There is no objectivity when I can deny you your work if you do not comply to my demands. Let's not even get into the morality, ethics, and integrity argument. redface.gif
Edited by dejanh - 12/6/12 at 11:04am
Etch-a-Sketch
(16 items)
 
  
CPUMotherboardGraphicsRAM
Intel Core i7 3930K Gigabyte GA-X79S-UP5 2x MSI R5850 OC 1.350V/1060/1260 in CrossFireX 64GB Corsair Dominator 1866MHz, 9-10-9-27-180-1T 
Hard DriveCoolingMonitorKeyboard
2x Corsair Force 120GB GT SSD in RAID-0, 2x WD ... EK Supreme HF (EN), HW Labs BI GTX480, HW Labs ... Samsung S27A850 Logitech G19 Gaming Keyboard 
PowerCaseMouseAudio
Corsair Professional Series Gold CMPSU-1200AX Antec 1200 Gamer Case, Custom Full-window Side ... Logitech G9 Laser Mouse Auzentech X-Fi Forte 
Other
Bigfoot Networks Killer Xeno Pro 
  hide details  
Reply
Etch-a-Sketch
(16 items)
 
  
CPUMotherboardGraphicsRAM
Intel Core i7 3930K Gigabyte GA-X79S-UP5 2x MSI R5850 OC 1.350V/1060/1260 in CrossFireX 64GB Corsair Dominator 1866MHz, 9-10-9-27-180-1T 
Hard DriveCoolingMonitorKeyboard
2x Corsair Force 120GB GT SSD in RAID-0, 2x WD ... EK Supreme HF (EN), HW Labs BI GTX480, HW Labs ... Samsung S27A850 Logitech G19 Gaming Keyboard 
PowerCaseMouseAudio
Corsair Professional Series Gold CMPSU-1200AX Antec 1200 Gamer Case, Custom Full-window Side ... Logitech G9 Laser Mouse Auzentech X-Fi Forte 
Other
Bigfoot Networks Killer Xeno Pro 
  hide details  
Reply
post #150 of 367
I just had a quick read of the last few pages and wanted to clear a few things up.

Intel documents a method to calculate CPU power consumption which all software uses. The sole purpose of this Watt power consumption number that the CPU calculates is for internal use so the CPU can control the Turbo Boost feature. It is based on VID voltage and has nothing to do with actual voltage. It can end up being a meaningless number if the core voltage you set in the bios is vastly different than the VID voltage that the CPU is requesting. Thanks to Dufus for pointing this out to me.

ThrottleStop reports VID voltage which is the voltage the CPU is requesting. Actual voltage on a desktop board is usually controlled by what you set in the bios.

It's impossible to determine actual amperage or power consumption based on this data. The only way to properly measure that is to tap into the CPU socket with the appropriate test equipment and measure it directly. In theory, a low VID processor should be able to overclock slightly better on this board compared to a processor with a high VID before hitting the throttling limit.

ThrottleStop only correctly supports the 4 core CPUs but in the screen shots I have seen, the data looks correct. It will only show 8 of the 12 threads though. There is another program in the RealTemp download called i7 Turbo GT. You can try running that to see if it shows all 12 threads. I think it should. Using high performance timers within the CPU, i7 Turbo GT and ThrottleStop can detect the first hints of CPU throttling when under load.

RealTemp 3.70
http://www.techpowerup.com/downloads/2089/Real_Temp_3.70.html

Has anyone tried to disable BD PROCHOT in ThrottleStop? This is the bi-directional processor hot signal path. It allows other items on the motherboard to tell the CPU that it is too hot which forces it to throttle. I don't think this is being used to throttle the CPU but it is a good idea to turn this off anyhow when testing.

The media is a farce. No one does any proper testing and when major problems are found, they are usually ignored. Every site is afraid to say anything negative for fear their supply of CPUs or motherboards or whatever is cut off.

The Asus P6T series had this exact same problem. Asus was nice enough to release some special bios versions for overclockers but they never released the fix on their website to the general public.
Edited by unclewebb - 12/6/12 at 1:42pm
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Intel Motherboards
Overclock.net › Forums › Intel › Intel Motherboards › Review of GIGABYTE's 2nd Gen. X79 Workstation MB: X79S-UP5