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 9

post #81 of 367
Quote:
Originally Posted by Sin0822 View Post

Then why doesn't ASUS have it?
That is my question, they are all about memory overclocking, yet they forgot tRC? That is what i am pointing out.
Asus is sadly known for using either non-standard naming for features, or for outright excluding essential features. I see your point, it's just that it does not hold up.
Quote:
Originally Posted by tsm106 View Post

??
tRC = tRAS + tRP
To my point above, they are setting it implicitly. Heck, for all you know they could be setting it as tRAS + tRP + n, where n is some "Asus defined" constant that they feel they should use to increment the minimum tRC value for stability.

In fact, go see post #5 and post #6 here http://www.xtremesystems.org/forums/showthread.php?280418-ASUS-Z77-Motherboards-Official-Support-Thread&p=5087599&viewfull=1#post5087599

What I am saying is exactly what Asus is doing, and the fact that they do not let you control the tRC, well that's a limitation of their boards. The last thing we want to do is start to define ourselves by shortcomings of others. Just because Asus made a lousy decision not to implement direct control of tRC, does not make it right, in fact it's completely wrong. The way they get around the problem is that they make tRC arbitrarily high, high enough that it will well exceed tRAS + tRP. They only way then to "control" the tRC is to keep bumping tRAS and/or tRP higher which sucks big time for memory performance.

Bottom line is - it's not the way to do it. Proper implementation for direct control of tRC values is essential for effective memory overclocking.
Edited by dejanh - 11/5/12 at 2:32pm
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 #82 of 367
The UP5 has manual tRC control, and it does appear to work, so evidently the platform does support tRC adjustment independent of tRC + tRAS.

The issue is that the range is needlessly limited to 32, which is too low for most anything past DDR3-1600 with all DIMMs populated.
Quote:
Originally Posted by tsm106 View Post

??
tRC = tRAS + tRP

Not always. tRP + tRAS is simply the default.

Many XMP profiles don't use this formula (I have a set of DDR3-2000 Patriot that defaults to tRP 9, tRAS 27, tRC 50), and it's not ideal for all memory configurations (the more ram I add, the more likely I am to need looser tRC than tRP + tRAS).

As 32GiB is absolutely the least amount of memory I'd consider putting in an LGA-2011 board, I can certainly make use of tRC control.
Primary
(15 items)
 
Secondary
(13 items)
 
In progress
(10 items)
 
CPUMotherboardGraphicsRAM
5820K @ 4.3/3.6GHz core/uncore, 1.225/1.2v Gigabyte X99 SOC Champion (F22n) 2x Sapphire R9 290X Tri-X OC New Edition (10036... 4x4GiB Crucial @ 2667, 12-11-12-27-T1, 1.37v 
Hard DriveHard DriveHard DriveCooling
Plextor M6e 128GB (fw 1.05) M.2 (PCI-E 2.0 2x) 2x Crucial M4 256GB 4x WD Scorpio Black 500GB Cooler Master Nepton 280L 
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 Reference R9 290X w/Stilt's MLU 1000e / 1375m E... 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 @ 4.3/3.6GHz core/uncore, 1.225/1.2v Gigabyte X99 SOC Champion (F22n) 2x Sapphire R9 290X Tri-X OC New Edition (10036... 4x4GiB Crucial @ 2667, 12-11-12-27-T1, 1.37v 
Hard DriveHard DriveHard DriveCooling
Plextor M6e 128GB (fw 1.05) M.2 (PCI-E 2.0 2x) 2x Crucial M4 256GB 4x WD Scorpio Black 500GB Cooler Master Nepton 280L 
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 Reference R9 290X w/Stilt's MLU 1000e / 1375m E... 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 #83 of 367
^ exactly my point...

Apologies for the wall of text below..I had to put it all in one post as it is important information and outlines exactly how to reproduce the throttling problem.

Going back to throttling now, Gigabyte got back to me as well. This was their response to me...
Quote:
Dear Customer,

We did not encounter any throttling issue when testing in our lab, you can try testing with the attached beta bios and check if it'll make any difference.

We’ve run the test with Prime95 v27.7 for 30 mins, the CPU frequency stayed at 4.2G and didn’t downclock at all (see attached image).

1. Check if the CPU frequency downclocking is caused by high CPU temperature.
2. Add a system fan to dissipate heat of MOSFET.

System Configuration:
CPU I7-3930K
M/B X79S-UP5-WIFI
BIOS F4l (CPU multiplier: 42, Vcore:1.2V)
RAM ADATA DDR3 1600 4G
VGA HD6950
HDD WD 250GB
OS WIN7 64bit
Cooling Corsair H100

However, I have responded to them noting that their test was much too conservative and that they are right, the setup that they specified above does not throttle in Prime95. I reproduced their configuration on my end (minus the video card) and indeed the chip did not throttle. However, going to 4.4GHz with a few minor tweaks did throttle again with the same setup. Therefore, I send them back the following message...
Quote:
Hi there,

I'm afraid that the test that was done was not sufficient to diagnose the problem. I have noted in one of my followup messages that there are certain conditions that exacerbate the situation (or in other words make throttling more likely to happen and happen more frequently). First, I have noticed that to really see throttling occuring the overclock has to be set to a minimum of 4.4GHz. Next, if you run LinX instead of Prime95 version 27.7 it is much easier to reproduce throttling at 4.4GHz. You can download an AVX compatible LinX version from the following link https://www.box.com/s/2pucmw2e6zcqojemnlol

If you set the problem size in LinX to 25,000 (using 4788 MB of memory) and then run the benchmark for 20-30 passes and observe the clock speed in CPU-Z you will notice the multiplier starting to change rapidly. In addition to this, I have noted a relationship between the memory frequency and the throttling occuring. Throttling occurs much less if the memory frequency is set to a low value, such as 1333MHz. However, if you increase the memory frequency to 1866MHz the throttling starts to happen a lot more and a lot more severely. Also, the throttling occurs in moments and then it goes away, and then it starts again. For instance, in 3-4 seconds the CPU will go from 44x, to 42x, to 41x, then back to 44x and so on. Therefore, I need to ask you to repeat the test and confirm whether issues occur with the following setup.

1. Set the CPU overclock to 4.4GHz by setting the multiplier to 44x.
2. Set the memory clock to 1866MHz and use at least 16GB of memory.
3. Leave all other settings at Auto, including CPU voltage.
4. Boot into Windows and start up CPU-Z and Prime95 version 27.7, then run the Blend stress test and observe continuously if the multiplier is changing in CPU-Z (usually starts around second pass of Prime95 Blend test).
5. After running Prime95, run the attached copy of LinX using a problem size of 25,000 (twenty five thousand) and/or memory size of 4788 MB (equivalent to problem size 25,000) for minimum of 20 passes. Once you press "Start" in LinX please note if the multiplier is changing.

I have also confirmed that my VRM temperatures are once again normal:
- measured on the front of the board at the base of the heatsink the temperatures were in the low 30C (approx 33C)
- measured on the back of the board at the heat-spreader (metal plate) the temperatures were in the high 30C (approx. 38C)

The CPU temperature did not exceed 70C on the hottest even during extended periods of running LinX at 25C ambient temperature. The average CPU core temperature is approximately 60C after extended runs at 4.4GHz using approximately 1.35V on the CPU.

I hope that this helps further in diagnosing the issue. I am looking forward to your response very soon.

I have also attached screenshots of the throttling with the setup I mentioned above to this message. Two screenshots are done at 4.4GHz using 1333MHz memory and the other two are done at 4.4GHz using 1866MHz memory.

Please advise as soon as you have repeated the tests. I will update the threads on Tweaktown and Overclock.net accordingly.

Sincerely,

Dejan

I have also attached the following screenshots to the message for them, showing the throttling in action...
LinX_Throttle_44x_Multi_1333MHz_DRAM.jpg 169k .jpg file
LinX_Throttle_44x_Multi_1866MHz_DRAM.jpg 175k .jpg file
Prime95_Throttle_44x_Multi_1333MHz_DRAM.jpg 174k .jpg file
Prime95_Throttle_44x_Multi_1866MHz_DRAM.jpg 131k .jpg file

@Sin0822 - If you have your setup still with this board, can you please try to redo the tests for throttling using the settings I outlined above. We're honestly losing our minds here...this is definitely happening and we want to help to have it resolved ASAP.
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 #84 of 367
Quote:
Originally Posted by $k1||z_r0k View Post

I just ordered the GA-X79S-UP5-WIFI... bloody can't wait to build a system with it. smile.gif
Any deals on SAS drives?
Well I bought one for $131 directly from Amazon with BT4/WiFi, so maybe they're already clearing these boards out?

did you get it?
I think it has been more than 1 week since, any one managed to get hold of a real up5s for $150?
    
CPUMotherboardGraphicsRAM
X4 940 Biostar 790GX 128M 4890 1GB 4GB 
Hard DriveOptical DriveOSMonitor
320GB 32X Samsung ODD Windows XP SP3 Dell FPW2407 
PowerCase
Corsair HX 520W Lian Li A07 
  hide details  
Reply
    
CPUMotherboardGraphicsRAM
X4 940 Biostar 790GX 128M 4890 1GB 4GB 
Hard DriveOptical DriveOSMonitor
320GB 32X Samsung ODD Windows XP SP3 Dell FPW2407 
PowerCase
Corsair HX 520W Lian Li A07 
  hide details  
Reply
post #85 of 367
Quote:
Originally Posted by dansi View Post

did you get it?
I think it has been more than 1 week since, any one managed to get hold of a real up5s for $150?

The wrong product was listed. I tried the same and got a refund the next day.
Quote:
Our system incorrectly identified a reconditioned model as a new unit and it was this listing that you purchased from. Our quality assurance team identified this while your order was being processed, and has asked us to cancel your order because we would never knowingly ship anything other than what is represented in our listings.

Because shipping labels had been generated for this order and transmitted to Amazon, I will be refunding your order in full at this time.

We apologize for any inconvenience as we sincerely appreciate your business.
post #86 of 367
Alright, so I have to provide this update...

Gigabyte got back to me again after performing the exact test that I have specified for them, even using the version of LinX that I have supplied. They are not able to reproduce any throttling with their setup and their Core i7 3930K C2 chip. I am really stumped as to what is going on here and why I am getting consistent throttling but they are not. They have even reproduced the test with Windows 8 at 4.4GHz using both Prime95 27.7 and LinX 0.6.4 with updated libraries for AVX support with the same result - no throttling. I am going to buy a different chip tonight to test out. If it does not resolve the problem then I am going to have to have the board itself replaced as there is really no option left then. I have already eliminated the DRAM as a possibility, and while the chip tested OK in another board, it could very well be that there is something wrong with it on this particular board.

More updates to come.

@Blameless - any chance that you can pick up a different chip to test as well in the same board?
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 #87 of 367
Gigabyte's sample could well be lower leakage/lower power than ours. They were also unable to produce throttling at my settings.

Individual parts of the same model often vary wildly in power consumption. For example, I've seen 130w TDP LGA-1366 consume as little as 60w under maximum load, while another chip of the same socket, stepping, cores, clock speeds, and TDP was pulling double that running the exact same test. This is an extreme case, but even 25% variation is fairly common.

It's not especially practical for me to test another chip in this board right now. The other board I borrowed is no longer available, and I don't have anyone locally who has a spare 2011 CPU they are likely to allow me to test. I'm also not keen on buying another CPU myself just to test with. Even if I did get another chip, and it did not throttle, that hardly implies anything is wrong with my original chip. I have tested the chip in another board (an acquaintance's ASUS P9X79), where it did not throttle, so it's obviously possible for the part to function without throttling.

At this point, I'd rather just get a different board, as that's where the problem lies. I'm not going to shop around for a lower power 3930K when it shouldn't matter how much current the thing draws. I'd rather have the board explode in a ball of fire than throttle without me telling it to. So, I think I'm just going to find a board that allows me to actually change or disable the TDP/OCP limits.
Edited by Blameless - 11/7/12 at 10:12am
Primary
(15 items)
 
Secondary
(13 items)
 
In progress
(10 items)
 
CPUMotherboardGraphicsRAM
5820K @ 4.3/3.6GHz core/uncore, 1.225/1.2v Gigabyte X99 SOC Champion (F22n) 2x Sapphire R9 290X Tri-X OC New Edition (10036... 4x4GiB Crucial @ 2667, 12-11-12-27-T1, 1.37v 
Hard DriveHard DriveHard DriveCooling
Plextor M6e 128GB (fw 1.05) M.2 (PCI-E 2.0 2x) 2x Crucial M4 256GB 4x WD Scorpio Black 500GB Cooler Master Nepton 280L 
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 Reference R9 290X w/Stilt's MLU 1000e / 1375m E... 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 @ 4.3/3.6GHz core/uncore, 1.225/1.2v Gigabyte X99 SOC Champion (F22n) 2x Sapphire R9 290X Tri-X OC New Edition (10036... 4x4GiB Crucial @ 2667, 12-11-12-27-T1, 1.37v 
Hard DriveHard DriveHard DriveCooling
Plextor M6e 128GB (fw 1.05) M.2 (PCI-E 2.0 2x) 2x Crucial M4 256GB 4x WD Scorpio Black 500GB Cooler Master Nepton 280L 
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 Reference R9 290X w/Stilt's MLU 1000e / 1375m E... 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 #88 of 367
Quote:
Originally Posted by Blameless View Post

Gigabyte's sample could well be lower leakage/lower power than ours. They were also unable to produce throttling at my settings.
Individual parts of the same model often vary wildly in power consumption. For example, I've seen 130w TDP LGA-1366 consume as little as 60w under maximum load, while another chip of the same socket, stepping, cores, clock speeds, and TDP was pulling double that running the exact same test. This is an extreme case, but even 25% variation is fairly common.
It's not especially practical for me to test another chip in this board right now. The other board I borrowed is no longer available, and I don't have anyone locally who has a spare 2011 CPU they are likely to allow me to test. I'm also not keen on buying another CPU myself just to test with. Even if I did get another chip, and it did not throttle, that hardly implies anything is wrong with my original chip. I have tested the chip in another board (an acquaintance's ASUS P9X79), where it did not throttle, so it's obviously possible for the part to function without throttling.
At this point, I'd rather just get a different board, as that's where the problem lies. I'm not going to shop around for a lower power 3930K when it shouldn't matter how much current the thing draws. I'd rather have the board explode in a ball of fire than throttle without me telling it to. So, I think I'm just going to find a board that allows me to actually change or disable the TDP/OCP limits.
I could not agree more...but I have to say I feel extremely disheartened by the whole experience. Do I just give up now or do I keep pushing to have this resolved? If I keep pushing, what do I do? How do I go about this to show that the board is the problem? I am going to try another chip, then flip it on the marketplace but the only other way I feel we can go about proving that there is an issue is to get another board all together and then run two tests sending the results of both to Gigabyte.

@Sin0822 - is there any way that you can get them to validate their wattage and amperage limits, as well as any built-in turbo TDP limitations?
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 #89 of 367
Another update...after some thinking and discussion with Blameless, I think that they are in fact seeing throttling happening too, just don't know what to look for...

Here is a screenshot that they sent me with LinX running that I marked up to point out some inconsistencies...



And here is my subsequent message to them...
Quote:
I'm writing this in relation to Email ID 1285616, a case that is already open where I reported the motherboard causing throttling of my chip when the processor is overclocked despite setting all power and current limits very high and having temperatures under control. I just realized a number of important things after analyzing your original screenshots and the newest screenshots you sent.

1. I noticed that your chip seems to be of an above-average quality when it comes to power requirements for a certain overclock. Different chips will have wildly different power curves, and in extreme cases two chips could be consuming totally different amounts of power, in fact on the order of multiples. For example, a very good low-leakage chip could consume only 60W-80W at a certain overclock and another chip at the same overclock could be consuming as much as 120W-140W (please note that I am not talking strictly about i7 3930K chips, just giving an example). I realize that this is an extreme example, but it happens and judging by your screenshots you are definitely dealing with a low-leakage chip since you required only 1.2V roughly for 4.2GHz, and less than 1.3V for 4.4GHz.

2. In the LinX screenshots that were taken, after close observation I have actually noticed that there are several instances where the GFlops drop significantly. This has to be result of CPU throttling in the back end, even if there was momentarily no observable change on the front end (that is no visible drop in CPU speed in CPU-Z). I have attached the screenshot that you have sent to me to this message and marked up the three GFlops results that are substantially lower than the others in the same run indicating throttling is occuring.

3. Can you please tell me if the board (by design) has a TDP/current limit, both for the VRM and for the CPU?

4. Related to (3) above, what is/are those TDP/current limits set to by default on the board? Based on what I have observed, I suspect that they are somewhere around 200W/130A.

5. Can you confirm that *all* of the UEFI BIOS features on the board that allow overriding the TDP/power and current limits for the CPU, VRM, etc. are working correctly? I especially want to see a verification that the VRM power and current limits are working correctly. In my tests, changing these values has no effect on anything on the board.

I am looking forward to your response. In the meantime, I will test my motherboard again with a different Core i7 3930K C2 CPU chosen at random from a different batch and a different location.

Please let me know your answers as soon as possible.

Thanks,
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 #90 of 367
So, I bought a new chip today, chosen at random from a completely different batch...here are the results

First, the new F4o BIOS made no difference at all...I am not sure what the change was supposed to be, but I cannot see any difference at all between this and F4n other than the fact that Gigabyte updated the RAID option ROMs

1. Throttling - No improvement, chip is still throttling all the time
2. Multiplier not set correctly when turbo is disabled - No improvement, disabling turbo and then setting a multiplier to any number always results in default multiplier (32x) once Windows loads
3. Memory compatibility - No improvement, still unable to run 8 banks for 2133MHz memory

Now, as I said, I even went out and bought a completely different chip, and the results are exactly the same...nay, they are even worse with the new chip. The throttling is even harder than before. The screenshots posted in this thread so far have been with a batch #3203C288 chip. The two screenshots below are done with a brand new chip I bought today, batch #3152C721. Throttling is even worse than before. This is also using BIOS F4o, just for the record. This is supposed to be a 4.4GHz overclock, with multiplier set at 44x.

Throttling all the way down to 32x...

Prime95_Throttle_44x_Multi_1866MHz_DRAM_New_Chip.jpg

And bouncing around, here at 39x

Prime95_Throttle_44x_Multi_1866MHz_DRAM_New_Chip_2.jpg

So, this is now multiple kits of memory tested at various speeds (from 1333MHz all the way up to 2133MHz) with no improvement to the throttling, as well as two completely different i7 3930K C2 chips chosen at random from two completely separate batches both throttling. Now either the firmware is seriously faulty, or I have a seriously faulty board. I will keep both chips until this is resolved, just to prove a point. There is a really serious problem with this board.
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
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