Overclock.net › Forums › Intel › Intel - General › New testing tool for Intel i7
New Posts  All Forums:Forum Nav:

New testing tool for Intel i7

post #1 of 23
Thread Starter 
Hello, I'm developing this new test tools (for my thesis) to verify the stability of the CPU Intel Core i7, the test is very fast, takes about 1 minute.

This test was developed specifically for a core i7 950, and has excellent results, it detects instability before and faster than IBT on the processor on which it was developed. I would like to see how it behaves on similar processors (tests carried out on the forum hwupgrade.it are giving good feedback http://www.hwupgrade.it/forum/showthread.php?t=2319365).

You'll do me a big favor if you can run some tests and maybe compare it with other programs such as first, IBT, Linx.
The procedure on a stable processor might be: slowly lowering the v-core, or increase the frequency of the CPU in order to bring in a situation of possible instability, and see if and at what point the test fails, and finds if fail first my test or LinX (or IBT)

Unlike the traditional stress test, this is intended to highlight the processor Speedpaths, which means the slowest path through the logic circuits of the CPU, which limits the maximum frequency, so does not need a running too long, and is very depended to processor on which it was developed.

The test is focused on the CPU, generates very few cache misses: ram-> cache (so the memory footprint of the program are a few MB instead of GB IBT & co.), Then the ram is not very tested, and temperatures cpu will not grow much.

Download quadcore version with HT (8 threads):http://www.cad.polito.it/research/Ev...es/fisa02c.zip

On this page you will find information about the project and the next versions of the tool: http://www.cad.polito.it/research/Ev...rclocking.html
The test may require Microsoft Visual C + + 2008 Redistributable Package or Microsoft Visual C + + 2010 Redistributable Package:
http://www.microsoft.com/downloads/e...3-37bf0912db84

Thanks!
post #2 of 23
Intel i7 930 @ 4.2ghz: i know that this setting is rock solid stable (tested for 150 ITB laps and 48 hours Prime95) and that if i bump the voltage down one notch then its unstable and wont pass 10 laps of ITB or 12 hours or Prime95.

knowing this, i tried out your program and it does load up the cores but the temps are not going very high when compared to ITB or Prime95 so it makes me wonder if the cores really are being stressed enough? id expect from a program that can detect instability in a few minutes a much higher load temp.

maybe you could increase or add a stress level? this program might be fine for a stock i7 950 but when overclocked the cores dont work as hard to do the same amount of work, and i find they need to be pushed harder the higher your clocks get.

overall though i am very interested in this program and would love to see it develop
Finally Finished!
(14 items)
 
  
CPUMotherboardGraphicsRAM
Intel i7 3930K Batch: 3135B495T Gigabyte X79 UD7 GTX 570 3-way SLi Corsair Vengeance 1600mhz CL9 
Hard DriveCoolingOSMonitor
60GB OCZ Vertex II / 3x 500GB Samsung Spinpoint EK Supreme HF Nickel, EK DCP4.0, Stealth GT 360... Windows 7 Ult 64bit Acer 22" G225HQ 
KeyboardPowerCaseMouse
Razor Lacosa Mirror Edition Silverstone Strider 1500w  Aerocool X-Preditor  CM Storm Sentinel Advanced 
Mouse Pad
Razer Goliathus Speed edition 
  hide details  
Reply
Finally Finished!
(14 items)
 
  
CPUMotherboardGraphicsRAM
Intel i7 3930K Batch: 3135B495T Gigabyte X79 UD7 GTX 570 3-way SLi Corsair Vengeance 1600mhz CL9 
Hard DriveCoolingOSMonitor
60GB OCZ Vertex II / 3x 500GB Samsung Spinpoint EK Supreme HF Nickel, EK DCP4.0, Stealth GT 360... Windows 7 Ult 64bit Acer 22" G225HQ 
KeyboardPowerCaseMouse
Razor Lacosa Mirror Edition Silverstone Strider 1500w  Aerocool X-Preditor  CM Storm Sentinel Advanced 
Mouse Pad
Razer Goliathus Speed edition 
  hide details  
Reply
post #3 of 23
Thread Starter 
Thank you testing! The characteristic of not increasing the temperature too much is due to the fact that the program works with few data only inside the cpu. Is built with a set of assembly instructions that move data within the CPU through various functional units, and try to detect a Speedpath, an error caused by too slow switching times trough this path.
It is a completely different approach from the classical stress test, where large amounts of data are moved between RAM and CPU cache. (this causes increase in temperature)

So to answer your question if the CPU is stressed enough:
this may be a good test for those who have no cooling problem, but I'm not sure if this same test is usable by different cpu (for now i had good feedback from some 920 and 950)

I'd glad if you can do a run of my test in the unstable condition where you don't pass 10 laps of ITB or 12 hours or Prime95.
So i can see if i'm at least at the level of IBT and PRIME (but more quick if fail) with your cpu.
Thanks
Edited by GianFisa - 3/14/11 at 8:27am
post #4 of 23
Adding unnecessary stress/heat doesn't make the program any better like liinx and p95.
I find folding to be the stability tool, many have seen 24 hour p95's 100+ linx pass ect. fail at folding and it runs about 4 -5 c less then p95 and 10c less then linx.

Looking forward to testing this out
post #5 of 23
Hey.

I have tested these on my E8400:
fisa01A
fisa02A
fisa02C

All 3 lets me pass, fisa01A is the fastest and lowest core temp.

Fisa02A and fisa02C lets me pass, but takes longer and around 3c higher core temps.

The thing is my system is NOT stable atm. My core#0 is failing after 30min p95 core#1 goes on forever.

I can pass IBT with some strange results.

But I can see potential in the program, with a little work it could be a very nice test
Edited by TwistedDane - 3/14/11 at 10:59am
    
CPUMotherboardGraphicsRAM
E8400 C0 @ 4.2(24/7 stable) P5E-WS Pro > Maximus Formula mod XFX HD6850 Black Edition 999/1150 4x1gb Corsair Dominator 
Hard DriveOSMonitorPower
Samsung F1 Windows7 x64 Ultimate Philips 221EL NorthQ Black Magic Flex 850w(50+A) 
CaseMouseMouse Pad
CoolerMaster HAF-932 Logitech MX-1000 Yes 
  hide details  
Reply
    
CPUMotherboardGraphicsRAM
E8400 C0 @ 4.2(24/7 stable) P5E-WS Pro > Maximus Formula mod XFX HD6850 Black Edition 999/1150 4x1gb Corsair Dominator 
Hard DriveOSMonitorPower
Samsung F1 Windows7 x64 Ultimate Philips 221EL NorthQ Black Magic Flex 850w(50+A) 
CaseMouseMouse Pad
CoolerMaster HAF-932 Logitech MX-1000 Yes 
  hide details  
Reply
post #6 of 23
Thread Starter 
Quote:
Originally Posted by TwistedDane View Post
Hey.

I have tested these on my E8400:
fisa01A
fisa02A
fisa02C

All 3 lets me pass, fisa01A is the fastest and lowest core temp.

Fisa02A and fisa02C lets me pass, but takes longer and around 3c higher core temps.

The thing is my system is NOT stable atm. My core#0 is failing after 30min p95 core#1 goes on forever.

I can pass IBT with some strange results.

But I can see potential in the program, with a little work it could be a very nice test
Hi, ty for test, the test 02* are developed for i7 microarchitecture, sure won't give good result with core microarchitecture, maybe the 01 could be good for you but is not optimized like the 02c.
post #7 of 23
Quote:
Originally Posted by GianFisa View Post
Thank you testing! The characteristic of not increasing the temperature too much is due to the fact that the program works with few data only inside the cpu. Is built with a set of assembly instructions that move data within the CPU through various functional units, and try to detect a Speedpath, an error caused by too slow switching times trough this path.
so what your saying is that this test is limited to the cpu only? That would be great to check if instability was caused by the cpu or somthing else. Unlike prime95 were you need to use the BSOD codes to try and identify the issue

Ill run some more tests for you tonight with some unstable settings. Also would you be interested in knowing if ram stability does create any impact on the test? I know you say that it is cpu based and only transfers between the core cache but have you tested for other factors like unstable ram?
Edited by Kal777 - 3/14/11 at 4:12pm
Finally Finished!
(14 items)
 
  
CPUMotherboardGraphicsRAM
Intel i7 3930K Batch: 3135B495T Gigabyte X79 UD7 GTX 570 3-way SLi Corsair Vengeance 1600mhz CL9 
Hard DriveCoolingOSMonitor
60GB OCZ Vertex II / 3x 500GB Samsung Spinpoint EK Supreme HF Nickel, EK DCP4.0, Stealth GT 360... Windows 7 Ult 64bit Acer 22" G225HQ 
KeyboardPowerCaseMouse
Razor Lacosa Mirror Edition Silverstone Strider 1500w  Aerocool X-Preditor  CM Storm Sentinel Advanced 
Mouse Pad
Razer Goliathus Speed edition 
  hide details  
Reply
Finally Finished!
(14 items)
 
  
CPUMotherboardGraphicsRAM
Intel i7 3930K Batch: 3135B495T Gigabyte X79 UD7 GTX 570 3-way SLi Corsair Vengeance 1600mhz CL9 
Hard DriveCoolingOSMonitor
60GB OCZ Vertex II / 3x 500GB Samsung Spinpoint EK Supreme HF Nickel, EK DCP4.0, Stealth GT 360... Windows 7 Ult 64bit Acer 22" G225HQ 
KeyboardPowerCaseMouse
Razor Lacosa Mirror Edition Silverstone Strider 1500w  Aerocool X-Preditor  CM Storm Sentinel Advanced 
Mouse Pad
Razer Goliathus Speed edition 
  hide details  
Reply
post #8 of 23
Is this only for quad cores?

Will testing it on my 980X give you any usefull information?
My System
(14 items)
 
  
CPUMotherboardGraphicsRAM
4770K @ 4.7GHz / 4400MHz Asus Sabertooth Z87 Gigabyte GTX 780 OC 8GB G.Skill Trident X 9-11-11-31-1T @ 2400MHz 
Hard DriveOptical DriveCoolingOS
OS SSD 840 Pro 256GB / FSX SSD 840 Pro 256GB LG DVD Thermalright SB-E x2 Win7 Ult 64b 
MonitorKeyboardPowerCase
Viewsonic VP2655wb 25.5" IPS LCD Logitech Wireless EX1110 w/Mouse Corsair AX860i CM HAF-X 
  hide details  
Reply
My System
(14 items)
 
  
CPUMotherboardGraphicsRAM
4770K @ 4.7GHz / 4400MHz Asus Sabertooth Z87 Gigabyte GTX 780 OC 8GB G.Skill Trident X 9-11-11-31-1T @ 2400MHz 
Hard DriveOptical DriveCoolingOS
OS SSD 840 Pro 256GB / FSX SSD 840 Pro 256GB LG DVD Thermalright SB-E x2 Win7 Ult 64b 
MonitorKeyboardPowerCase
Viewsonic VP2655wb 25.5" IPS LCD Logitech Wireless EX1110 w/Mouse Corsair AX860i CM HAF-X 
  hide details  
Reply
post #9 of 23
Programs that cause the temp to increase to crazy levels aren't really testing stability. They are good for testing your cooling, but IBT isn't the end all be all of stability.

I know people are always trying to find easy, quick ways to test stability, but the only way to actually make sure your system is fully stable is to run WCG 24/7. If you get no crashes and your results are without errors then you have a stable system.

I don't get all the people who overclock to these crazy high clocks and then don't actually do something worthwhile with their computing power.
Will of D
(18 items)
 
  
CPUMotherboardGraphicsRAM
Intel 980x EVGA E759 Classified LE 2x EVGA 580SC 12GB Corsair GT 
Hard DriveHard DriveHard DriveOptical Drive
Sandisk Extreme OCZ Vertex LE Cavier Black random 
CoolingOSMonitorKeyboard
Watercooled Win 7 x64 NEC 30", 2x NEC P221W's DAS keyboard 
PowerCaseMouseMouse Pad
Antec 1000w Custom Logitech G9 X-Trac 
AudioAudio
SB X-fi Titanium to Yamaha 5.1 receiver Audio-GD NB-12 DAC to Musical Paradise MP-301 a... 
  hide details  
Reply
Will of D
(18 items)
 
  
CPUMotherboardGraphicsRAM
Intel 980x EVGA E759 Classified LE 2x EVGA 580SC 12GB Corsair GT 
Hard DriveHard DriveHard DriveOptical Drive
Sandisk Extreme OCZ Vertex LE Cavier Black random 
CoolingOSMonitorKeyboard
Watercooled Win 7 x64 NEC 30", 2x NEC P221W's DAS keyboard 
PowerCaseMouseMouse Pad
Antec 1000w Custom Logitech G9 X-Trac 
AudioAudio
SB X-fi Titanium to Yamaha 5.1 receiver Audio-GD NB-12 DAC to Musical Paradise MP-301 a... 
  hide details  
Reply
post #10 of 23
Quote:
Originally Posted by GianFisa View Post
snip
So its written to test the stability of a processor on its most basic level (i.e. assembly code processing)? It tests what all the other processing types work off of? Thats BRILLIANT.
If you develop this all the way it will be far and away better than IBT or P95. And coupling it with one or both will be nearly bulletproof for stability testing.

And I agree. I only OC my processor to what I need.
Ive never seen a point in anything above ~3.7Ghz.
Micro Mule
(11 items)
 
  
CPUMotherboardGraphicsRAM
i7 6700k MSI Z170I Gaming Pro AC Gigabyte GTX 1070 G1 Gaming 16GB GSkill Ripjaws V DDR4 3200 
Hard DriveCoolingMonitorPower
250GB Crucial MX200, 960GB Adata , 1TB WD Black Noctua NH-L12 LG 27UD68 4k Seasonic M12II 620W 
CaseMouseAudio
Corsair Obsidian 250D Logitech Performance Mouse MX Topping TP-30 Mk2, 2x Polk M10, Sony MDR-7506 
  hide details  
Reply
Micro Mule
(11 items)
 
  
CPUMotherboardGraphicsRAM
i7 6700k MSI Z170I Gaming Pro AC Gigabyte GTX 1070 G1 Gaming 16GB GSkill Ripjaws V DDR4 3200 
Hard DriveCoolingMonitorPower
250GB Crucial MX200, 960GB Adata , 1TB WD Black Noctua NH-L12 LG 27UD68 4k Seasonic M12II 620W 
CaseMouseAudio
Corsair Obsidian 250D Logitech Performance Mouse MX Topping TP-30 Mk2, 2x Polk M10, Sony MDR-7506 
  hide details  
Reply
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Intel - General
Overclock.net › Forums › Intel › Intel - General › New testing tool for Intel i7