Overclock.net › Forums › AMD › AMD CPUs › [OFFICIAL] FX-8320/FX-8350 Vishera Owners Club
New Posts  All Forums:Forum Nav:

[OFFICIAL] FX-8320/FX-8350 Vishera Owners Club - Page 467

post #4661 of 67356
bmgjet, I have looked at the entire changelog (long). These are the only references to AVX.
Quote:
New features in Version 27.3 of prime95.exe

1) 64-bit optimizations for AVX-capable computers. Single core benchmarks on Intel Sandy
Bridge computers should see a decent speed increase. However, running on 4 cores
runs into main memory bandwidth bottleneck. Sandy Bridge-E CPUs with 4 memory channels
should avoid this problem.

New features in Version 27.1 of prime95.exe

1) 32-bit FFTs optimized for AVX-capable computers. Intel Sandy Bridge computers should
see a 25% speed increase.

There were no AVX bugs mentioned in the log...


m0bius, it is not that all other applications are broken. Not all of them do the same thing.
Edited by bios0110 - 11/30/12 at 10:50pm
post #4662 of 67356
I'll settle with 4.5GHz on this motherboard Don't want to stress it too hard, because of the 4+2 phase.
Oxygène
(15 items)
 
  
CPUMotherboardGraphicsRAM
Ryzen 5 1600 MSI B350 Tomahawk Arctic VTX3D R9 390 8GB G.Skill Ripjaws  
Hard DriveHard DriveHard DriveHard Drive
Kingston Hyper X Kingston HyperX Fury ADATA SP550 ADATA SX8000 
CoolingOSMonitorKeyboard
Noctua NH-D15 SE-AM4 Windows 10 Dell U2711 Microsoft WK 800 
PowerCaseMouse
Chieftec A-135 Series 1000W NZXT. Phantom 410 Microsoft WM 1000 
  hide details  
Reply
Oxygène
(15 items)
 
  
CPUMotherboardGraphicsRAM
Ryzen 5 1600 MSI B350 Tomahawk Arctic VTX3D R9 390 8GB G.Skill Ripjaws  
Hard DriveHard DriveHard DriveHard Drive
Kingston Hyper X Kingston HyperX Fury ADATA SP550 ADATA SX8000 
CoolingOSMonitorKeyboard
Noctua NH-D15 SE-AM4 Windows 10 Dell U2711 Microsoft WK 800 
PowerCaseMouse
Chieftec A-135 Series 1000W NZXT. Phantom 410 Microsoft WM 1000 
  hide details  
Reply
post #4663 of 67356
Quote:
Originally Posted by m0bius View Post

Sure they all work differently, but they aren't so different that affected sub systems are completely exclusive to one test.

I am not sure, it could just be different sequences of instructions that cause different issues.
Quote:
My problem with these arguments is that from my point of view, this means that all other benchmarks are completely and utterly useless, solely because Prime fails my OC.

Well, they are not necessarily completely useless, but running these other benchmarks alone doesn't seem to be an adequate test.
I haven't seen any case where IBT fails and Prime95 passes so far though.
Quote:
Usually when a problem like this comes up, logic dictates that since 1 out of 5 stress tests failed, the 1 is the exception and should be investigated. Now we're saying that because Prime failed all other stress tests are broken?

No, logic is failing you here.
Just because the other tests pass doesn't mean they are themselves broken. How could you infer that ?
The one exception should definitely be investigated, I agree with you. So far I have not seen any evidence that it's broken with Vishera in general. Just broken with Vishera OC.
Quote:
Prime frequently needs patching with new CPU releases, has it been patched with Vishera? (Honest question, I haven't seen or heard anything recently.)

No, it has not been patched for Vishera, the latest build is still pre-Vishera.

There is no obvious reason why it needs to be patched for Vishera. There could be optimizations to take advantage of new instructions. But that still doesn't mean the old version should necessarily fail.

Of course it could be a bug, but it would be a strange one that doesn't seem to manifest itself without OC.
Quote:
Also, there have been several people who are unable to pass Prime at stock.

Now, that is interesting, any links to those reports ?

My own run at stock settings with no power management is 6 hours, and not one error. If it goes past 7 hours it will be better than I have gotten with any OC. I will let it run up to 12 hours at least.
post #4664 of 67356
Quote:
Originally Posted by sdlvx View Post

Prime95 has problems, can we please drop this?

https://www.google.com/search?q=prime95+stock+crash&ie=utf-8&oe=utf-8&aq=t

Even Athlons have had problems running Prime95 before it got patched.

What SD said. I do not understand how anyone can say P95 does not have a problem period. It has patches and revisions for Intel processors, and patches in general. why would it be ruled out then that it may need a patch to work with PD? One of the patches was to incease performance with IB/SB after all.

Quote:
Originally Posted by kzone75 View Post

I'll settle with 4.5GHz on this motherboard Don't want to stress it too hard, because of the 4+2 phase.
Kz, do you have a fan on the VRM/NB? they get very warm on that board with heavy PD OC's




I am continuing to put together a comprehensive OC bench @ 5.0,5.1, and 5.2GHz.
Heaven 3.0 @ 5.0GHz settings maxed out, GPU's @ stock.


I just added a Nvidia GT 640 2GB to my 3 x HD 7970's for PhysX, Is anyone else doing this? If you are would you PM me?
Thanks smile.gif
Edited by Red1776 - 11/30/12 at 11:17pm
2.2 kW Holodeck
(14 items)
 
Chernobyl
(0 items)
Power Station
(0 items)
CPUMotherboardGraphicsRAM
AMD FX-8350 @ 5.2GHz Gigabyte GA-990-FXA-UD7 4 x MSI R 290X 4GB Gaming 16GB Corsair Dominator Platinum 2133 
Hard DriveOptical DriveCoolingOS
4 x Seagate 1 TB 7200 2 x Asus Optiarc Koolance 370 CPU WB/ AlphaCool Water cooling: ... Windows 7 Professional 
MonitorKeyboardPowerCase
1x3+1 Eyefiniy 5760x1080 Gigabyte 8100K V.2 1-AX1200w + 2x FSP x500w Coolermaster Cosmos II Full Tower 
  hide details  
Reply
2.2 kW Holodeck
(14 items)
 
Chernobyl
(0 items)
Power Station
(0 items)
CPUMotherboardGraphicsRAM
AMD FX-8350 @ 5.2GHz Gigabyte GA-990-FXA-UD7 4 x MSI R 290X 4GB Gaming 16GB Corsair Dominator Platinum 2133 
Hard DriveOptical DriveCoolingOS
4 x Seagate 1 TB 7200 2 x Asus Optiarc Koolance 370 CPU WB/ AlphaCool Water cooling: ... Windows 7 Professional 
MonitorKeyboardPowerCase
1x3+1 Eyefiniy 5760x1080 Gigabyte 8100K V.2 1-AX1200w + 2x FSP x500w Coolermaster Cosmos II Full Tower 
  hide details  
Reply
post #4665 of 67356
Thread Starter 
Quote:
Originally Posted by madbrain View Post

Quote:
Originally Posted by m0bius View Post

Sure they all work differently, but they aren't so different that affected sub systems are completely exclusive to one test.

I am not sure, it could just be different sequences of instructions that cause different issues.
Quote:
My problem with these arguments is that from my point of view, this means that all other benchmarks are completely and utterly useless, solely because Prime fails my OC.

Well, they are not necessarily completely useless, but running these other benchmarks alone doesn't seem to be an adequate test.
I haven't seen any case where IBT fails and Prime95 passes so far though.
Quote:
Usually when a problem like this comes up, logic dictates that since 1 out of 5 stress tests failed, the 1 is the exception and should be investigated. Now we're saying that because Prime failed all other stress tests are broken?

No, logic is failing you here.
Just because the other tests pass doesn't mean they are themselves broken. How could you infer that ?
The one exception should definitely be investigated, I agree with you. So far I have not seen any evidence that it's broken with Vishera in general. Just broken with Vishera OC.
Quote:
Prime frequently needs patching with new CPU releases, has it been patched with Vishera? (Honest question, I haven't seen or heard anything recently.)

No, it has not been patched for Vishera, the latest build is still pre-Vishera.

There is no obvious reason why it needs to be patched for Vishera. There could be optimizations to take advantage of new instructions. But that still doesn't mean the old version should necessarily fail.

Of course it could be a bug, but it would be a strange one that doesn't seem to manifest itself without OC.
Quote:
Also, there have been several people who are unable to pass Prime at stock.

Now, that is interesting, any links to those reports ?

My own run at stock settings with no power management is 6 hours, and not one error. If it goes past 7 hours it will be better than I have gotten with any OC. I will let it run up to 12 hours at least.

More importantly, has it been patched for Bulldozer. If you can't see why that could be an issue, I have some wonderful slides from AMD that show how the new design works. tongue.gif

There are a few in the thread, both for 8320s and 8350s. They are dozens of pages ago, when we had the last Prime war, so I'm hoping those people will speak up again. I'm not going searching.
Forge
(17 items)
 
Forge-LT
(7 items)
 
 
CPUMotherboardGraphicsGraphics
Intel i7-5960X (4.625Ghz) ASUS X99-DELUXE/U3.1 EVGA 1080ti SC2 Hybrid EVGA 1080ti SC2 Hybrid 
RAMHard DriveCoolingOS
64GB Corsair Dominator Platinum (3000Mhz 8x8GB) Samsung 950 Pro NVMe 512GB EK Predator 240 Windows 10 Enterprise x64 
MonitorKeyboardPowerCase
2x Acer XR341CK Corsair Vengeance K70 RGB Corsair AX1200 Corsair Graphite 780T 
MouseAudioAudioAudio
Corsair Vengeance M65 RGB Sennheiser HD700 Sound Blaster AE-5 Audio Technica AT4040 
Audio
30ART Mic Tube Amp 
CPUMotherboardGraphicsRAM
i7-4720HQ UX501JW-UB71T GTX 960m 16GB 1600 9-9-9-27 
Hard DriveOSMonitor
512GB PCI-e SSD Windows 10 Pro 4k IPS 
  hide details  
Reply
Forge
(17 items)
 
Forge-LT
(7 items)
 
 
CPUMotherboardGraphicsGraphics
Intel i7-5960X (4.625Ghz) ASUS X99-DELUXE/U3.1 EVGA 1080ti SC2 Hybrid EVGA 1080ti SC2 Hybrid 
RAMHard DriveCoolingOS
64GB Corsair Dominator Platinum (3000Mhz 8x8GB) Samsung 950 Pro NVMe 512GB EK Predator 240 Windows 10 Enterprise x64 
MonitorKeyboardPowerCase
2x Acer XR341CK Corsair Vengeance K70 RGB Corsair AX1200 Corsair Graphite 780T 
MouseAudioAudioAudio
Corsair Vengeance M65 RGB Sennheiser HD700 Sound Blaster AE-5 Audio Technica AT4040 
Audio
30ART Mic Tube Amp 
CPUMotherboardGraphicsRAM
i7-4720HQ UX501JW-UB71T GTX 960m 16GB 1600 9-9-9-27 
Hard DriveOSMonitor
512GB PCI-e SSD Windows 10 Pro 4k IPS 
  hide details  
Reply
post #4666 of 67356
Quote:
Originally Posted by bios0110 View Post

Some guy said that logic dictates that if only 1 application out of 5 fails consistently then there must be something wrong with the program.
Actually, that is not the case.
When a software engineer makes an application such as this, she/he must decide what to test and how to test it.

For all you know, most of the "test" applications you use don't actually test anything. That is, it could very well be that they just give the CPU some work to do and if that work finishes then you "passed". You have no idea if the computational results of each operation were even checked by the software.
Simplified example to show the point:
Think of these as series of operations in 1 run:
2 + 5 (= 7 answer ignored, not verified)
5 + 8 (= 20 answer ignored, not verified)
23 / 3 (= 7.5784 answer ignored, not verified)
45 / 5 (= 8.9999 answer ignored, not verified)
...
Congratulations! You passed. smile.gif (Because each operation completed without the app crashing)
A smarter test would instead verify the result of each operation:
2 + 5 = 7 (true)
5 + 8 = 20 (Wrong, programmer bothered to check the result, now must decide to terminate test or let it run)
23 / 3 = 7.5784 (Wrong, programmer bothered to check the result, now must decide to terminate test or let it run)
46 / 5 = 9.9999 (Wrong, programmer bothered to check the result, now must decide to terminate test or let it run)
Keeping this simple, as you can see, you have no idea what is actually done by these applications at even such a trivial level. Even if some operations are checked other operations may not be.
Adding to that, you don't know what actual operations are being performed. In the end, at the assembly level, it is all simple registry operations, which if you know anything about programming would know that these operations are CPU agnostic across x86 x86-64. In other words, they don't have to rewrite the app just because AMD release a new x86-64 CPU...
Prime95 does not have FX8320 issues because I have ran it for 12hrs with no errors.
Overclocked, for me, Prime95 stability is vcore related. With less vcore test fails sooner. So it is obvious that this is an overclock issue and not faulty code issue. The code doesn't become better code proportional to the amount of extra volts I add (test runs longer)... The code is fixed. With high volts the overclock is stable because the results of the operations performed by prime95 are correct and if I give it less volts the results are not correct. Simple as that.
I have read what the IBT guy said (that IBT can heat better and find stability issues faster than Prime95). In practice I don't find this to be true at all. It doesn't heat up my FX8320 or 990FXA mobo any better than Prime95. Nor does it ever find errors for me. Even when extremely unstable and Prime95 finds error in 3 min. Most other benchmarks barely heat up my FX8320. I've ran all the top of benchmarking apps and my scores are what they should be but these apps barely budge the stock temps. Prime95 maxes them (at 60 deg C).

Thank you for taking the time to write this - from another software engineer here.
Personally, I would rather have my system crash than gives me wrong answers and keep going.
Who knows what silent errors can lead to down the line, corrupt data on the hard disk, etc.

As an example, during one of my OC tests, when I booted, the primary display changed to the monitor on the left when I booted for no good reason. Normally, my primary display is the middle of my 3 monitors. Most likely, the 5 GHz OC on air caused that weird behavior .
Who knows what else could have been going wrong silently due to the OC, this was just after booting.

Most production programs do not do the type of error checking that could cause these conditions because they are not normally possible with a working CPU and RAM. Writing that code is expensive , and takes a lot of extra cycles to execute. Only debug versions of the program will have (some) of those checks, ie. assertions which are meant to detect program bugs, but could also detect problems caused by OC.
I bet just booting an OS with a debug kernel successfully would be a good starting point for testing OC.
post #4667 of 67356
Quote:
Originally Posted by bmgjet View Post

Have a look in the change log, They had to fix bugs for SB and IB using the AVX instruction set. PD has the same instruction set.

So maybe we need to run an older version of Prime95 without AVX, if the bug is in the AVX code .
post #4668 of 67356
Quote:
Originally Posted by m0bius View Post

I was that guy, and I said it should be investigated, not necessarily wrong.
I simply can't wrap my head around how all other tests are now suddenly broken, and apparently don't stress your system correctly allowing 7+ hrs of stable testing, all because Prime95 is infallable. I'm done with this topic (not out of anger/rage, just that this topic won't die.)

Who said the other tests were broken ?
The fact that they pass most of the time would point to them *not* being broken.
They just seem insufficient.
post #4669 of 67356
Thread Starter 
Quote:
Originally Posted by madbrain View Post

Quote:
Originally Posted by bmgjet View Post

Have a look in the change log, They had to fix bugs for SB and IB using the AVX instruction set. PD has the same instruction set.

So maybe we need to run an older version of Prime95 without AVX, if the bug is in the AVX code .

He's just suggesting that it needs patches to work correctly all the time on PD. Theoretically, Prime95 27.7 has the needed patch for the Bulldozer design.
Forge
(17 items)
 
Forge-LT
(7 items)
 
 
CPUMotherboardGraphicsGraphics
Intel i7-5960X (4.625Ghz) ASUS X99-DELUXE/U3.1 EVGA 1080ti SC2 Hybrid EVGA 1080ti SC2 Hybrid 
RAMHard DriveCoolingOS
64GB Corsair Dominator Platinum (3000Mhz 8x8GB) Samsung 950 Pro NVMe 512GB EK Predator 240 Windows 10 Enterprise x64 
MonitorKeyboardPowerCase
2x Acer XR341CK Corsair Vengeance K70 RGB Corsair AX1200 Corsair Graphite 780T 
MouseAudioAudioAudio
Corsair Vengeance M65 RGB Sennheiser HD700 Sound Blaster AE-5 Audio Technica AT4040 
Audio
30ART Mic Tube Amp 
CPUMotherboardGraphicsRAM
i7-4720HQ UX501JW-UB71T GTX 960m 16GB 1600 9-9-9-27 
Hard DriveOSMonitor
512GB PCI-e SSD Windows 10 Pro 4k IPS 
  hide details  
Reply
Forge
(17 items)
 
Forge-LT
(7 items)
 
 
CPUMotherboardGraphicsGraphics
Intel i7-5960X (4.625Ghz) ASUS X99-DELUXE/U3.1 EVGA 1080ti SC2 Hybrid EVGA 1080ti SC2 Hybrid 
RAMHard DriveCoolingOS
64GB Corsair Dominator Platinum (3000Mhz 8x8GB) Samsung 950 Pro NVMe 512GB EK Predator 240 Windows 10 Enterprise x64 
MonitorKeyboardPowerCase
2x Acer XR341CK Corsair Vengeance K70 RGB Corsair AX1200 Corsair Graphite 780T 
MouseAudioAudioAudio
Corsair Vengeance M65 RGB Sennheiser HD700 Sound Blaster AE-5 Audio Technica AT4040 
Audio
30ART Mic Tube Amp 
CPUMotherboardGraphicsRAM
i7-4720HQ UX501JW-UB71T GTX 960m 16GB 1600 9-9-9-27 
Hard DriveOSMonitor
512GB PCI-e SSD Windows 10 Pro 4k IPS 
  hide details  
Reply
post #4670 of 67356
Ahhh the thread that never sleeps.
In my case it was a combination of a very low vid chip ( 1.28) and no LLC that caused prime 95 to fail at stock. I watched the voltage and every time it dipped below 1.28 due to vdroop a worker would throw an error. I added .03 volts and it can prime longer than i have the patience to let it run. I repeated this experiment half a dozen times with the same results.
Ryzen Shine!
(14 items)
 
  
CPUMotherboardGraphicsRAM
Ryzen 1800X @ 4125mhz 1.432 Volts MSI X370 Titanium   Fury  G.SKILL TridentZ Series 32GB (4 x 8GB) 288-Pin... 
Hard DriveOptical DriveCoolingOS
Samsung 850 pro Hp 1260 I liquid unobtanium Koolance 480mm radiator, 39... Win 7 HP/Winspy 10 64bit 
MonitorKeyboardPowerCase
24" hp @ 1900x1200 Logitech G19 PC power and cooling 910 watt silencer Thermaltake P5 
MouseMouse Pad
Logitech MX 518 fUnc industries 
  hide details  
Reply
Ryzen Shine!
(14 items)
 
  
CPUMotherboardGraphicsRAM
Ryzen 1800X @ 4125mhz 1.432 Volts MSI X370 Titanium   Fury  G.SKILL TridentZ Series 32GB (4 x 8GB) 288-Pin... 
Hard DriveOptical DriveCoolingOS
Samsung 850 pro Hp 1260 I liquid unobtanium Koolance 480mm radiator, 39... Win 7 HP/Winspy 10 64bit 
MonitorKeyboardPowerCase
24" hp @ 1900x1200 Logitech G19 PC power and cooling 910 watt silencer Thermaltake P5 
MouseMouse Pad
Logitech MX 518 fUnc industries 
  hide details  
Reply
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: AMD CPUs
Overclock.net › Forums › AMD › AMD CPUs › [OFFICIAL] FX-8320/FX-8350 Vishera Owners Club