Awesome new CPU stress utility needs feedback!

[zzzzzzzzzz]

An error is returned when Core Damage 0.8f is run on a computer with a Palomino mobile AMD Athlon 4 processor ("problem").

I have attached an image of the error message and CPU-Z screenshot of a mobile AMD Athlon 4 processor for which Core Damage fails on.

Please have the problem corrected such that Core Damage runs without crashing.

EDIT: Toast runs on the machine with the Palomino core processor without issue.

 

[Nanerbeet]

Ahhh yes the latest version only works on processors with SSE-2 instructions. I didn't bother to add a check for SSE-2 support 'cuz I figured SSE-1 processors were extinct

 

[zzzzzzzzzz]

Ahhh yes the latest version only works on processors with SSE-2 instructions. I didn't bother to add a check for SSE-2 support 'cuz I figured SSE-1 processors were extinct

For my stress testings, I usually use Core Damage or Toast to test a CPU at maximum as feasible load.

I do know of computers in use that still use processors that do not have support for SSE2.

If Core Damage can be made to use Toast's stress algorithm for processors that do not support SSE2, it can completely replace Toast.

It may be good to have an updated portable version of an Core Damage available again (no installation).

 

[hankafyin]

Ouch! If I'm gonna run this I will need some water. OCCT linpack is nothing compared to this. My Q9550 said HELL NO !! I'm gonna try it on the 955. The 955 @ 4.0 on air (CM Hyper 212+ push/pull) topped out at 56c, about the same as OCCT linpack. I need better cooling for the intel rig.

 

[CompuTamer]

I just ran it, and my i7 hit 91C. Usually it tops out at 83ish under P95, so i'd say it's a bit hotter. It also heats up faster.

P95 will crash earlier than this though. After an hour of this, it didn't crash undervolted. P95 BSODed after 4 minutes :/

 

[infam0ussteven]

I think I should give this a try on my C2D rig to truly see if my undervolting is stable or not. Will comes back with results later on when I get home!

 

[RJARRRPCGP]

If you suspect the crash is not related to overclock, run it at stock speed, stock voltage just to verify.

If the processor overheats, it will hang or reboot. If the processor fails a calculation (such as memory address or branch mis-prediction), it will crash. Not exactly "built-in" error checking; its just that the elevated heat condition might help trigger a failure.

The reboot is when Windows detects an error. When the Windows kernel freaks out, it reboots the PC by default.

BSOD usually only occurs when a program don't have error handling.

That's why it's rare to get a BSOD when Prime95 fails.

When the processor fails a calculation, that can cause
STOP: 0x0000009C MACHINE_CHECK_EXCEPTION

or

STOP: 0x0000000A IRQL_NOT_LESS_OR_EQUAL

 

[SLeePYG72786]

I like it!

Here are my results after running it for only 5 minutes:

Intel Core 2 Duo E7200 (2.53GHz) @ 3.3GHz, 1.2102V

Everest Ultimate:
CPU > 64
Core 1 > 63
Core 2 > 64

It's about the same as I get for IntelBurnTest, which I use more than anything else. Good job man!

 

[zzzzzzzzzz]

I have tried using Core Damage 0.8h to test my Intel Core i7-870 with Turbo Boost enabled.

The frequency reported by Core Damage 0.8h is significantly less than what the processor is reported to be running at by CPU-Z 1.59.

It seems that Core Damage 0.8h reports the frequency as if Turbo Boost were not enabled, while CPU-Z reports actual frequency (with Turbo Boost enabled).

I think that the Core Damage 0.8h not (apparently) accounting for Intel Turbo Boost might be a bug/defect.