akosf C37 performance observations thread

nairb
nairb
Joined: 21 Feb 05
Posts: 22
Credit: 6335590
RAC: 219

I thought I might give the

I thought I might give the new opt app ago. Tried the C37 first. All 3 amd xp machines gave validate error. Switched to the A36 app and sofar still have the validate errors.
Will the optimised app only be ok with new wu?. All machines use win98 and have worked fine with the standard app.

Thanks.

Nairb

Akos Fekete
Akos Fekete
Joined: 13 Nov 05
Posts: 561
Credit: 4527270
RAC: 0

RE: Will the optimised app

Message 25977 in response to message 25976

Quote:
Will the optimised app only be ok with new wu?.


It should work with 'half' wus.
I did not find the solution of this failure yet.
Probably it is independent of optimisation.

Ziran
Ziran
Joined: 26 Nov 04
Posts: 194
Credit: 356403
RAC: 1626

RE: I thought I might give

Message 25978 in response to message 25976

Quote:

I thought I might give the new opt app ago. Tried the C37 first. All 3 amd xp machines gave validate error. Switched to the A36 app and sofar still have the validate errors.
Will the optimised app only be ok with new wu?. All machines use win98 and have worked fine with the standard app.


Could you pleas try version 387.
Also look at this tread .

Then you're really interested in a subject, there is no way to avoid it. You have to read the Manual.

nairb
nairb
Joined: 21 Feb 05
Posts: 22
Credit: 6335590
RAC: 219

Thanks for the info. I have

Thanks for the info. I have set them all back to the standard app for a while.
I will try one on the 387 app to see if it stops the validation errors.

Alinator
Alinator
Joined: 8 May 05
Posts: 927
Credit: 9352143
RAC: 0

So far I've had 2 failures on

So far I've had 2 failures on one K6-2 with C37, and 1 success on the other K6-2 and a second in progress.

The K6/300 has had 2 go through OK, 1 pending and one in progress, but they were shorties.

Alinator

Akos Fekete
Akos Fekete
Joined: 13 Nov 05
Posts: 561
Credit: 4527270
RAC: 0

RE: So far I've had 2

Message 25981 in response to message 25980

Quote:

So far I've had 2 failures on one K6-2 with C37, and 1 success on the other K6-2 and a second in progress.

The K6/300 has had 2 go through OK, 1 pending and one in progress, but they were shorties.

Alinator

Thanks for these infos!
I hoped that C37 will be much faster on your K6 machines...

I tried to find the reason of the validation problem.
I know that the precision of C37 has to be very near to the original.
I made a code that gave poor results, but the validator accepted them!
So, I examined the codes again and I saw there is a critical point.
Sometimes the exponent of the number format is not enough.
Probably this is the reason of the validation problem of SSE based linux application too.

edit:
exponent size of SSE -> 8 bit
exponent size of SSE2 -> 11 bit
exponent size of FPU -> 15 bit

LiborA
LiborA
Joined: 8 Dec 05
Posts: 74
Credit: 337135
RAC: 0

RE: RE: So far I've had 2

Message 25982 in response to message 25981

Quote:
Quote:

So far I've had 2 failures on one K6-2 with C37, and 1 success on the other K6-2 and a second in progress.

The K6/300 has had 2 go through OK, 1 pending and one in progress, but they were shorties.

Alinator

Thanks for these infos!
I hoped that C37 will be much faster on your K6 machines...

I tried to find the reason of the validation problem.
I know that the precision of C37 has to be very near to the original.
I made a code that gave poor results, but the validator accepted them!
So, I examined the codes again and I saw there is a critical point.
Sometimes the exponent of the number format is not enough.
Probably this is the reason of the validation problem of SSE based linux application too.

edit:
exponent size of SSE -> 8 bit
exponent size of SSE2 -> 11 bit
exponent size of FPU -> 15 bit

Do you recomend use C37 or not?

Akos Fekete
Akos Fekete
Joined: 13 Nov 05
Posts: 561
Credit: 4527270
RAC: 0

RE: Do you recomend use C37

Message 25983 in response to message 25982

Quote:
Do you recomend use C37 or not?

This is an unofficial application, so "use at your own risk".

C37 uses FPU (maximum precision) for the calculation, but it needs about 4 bit longer exponent than original.
I checked C37 on my computers with 70-80 units without any error.
A longer test period would be able to determine the frequency of the faults.

I hope that S38 will solve this exponent problem.
It consists a very aggressive precision combination method to extend the precision.

LiborA
LiborA
Joined: 8 Dec 05
Posts: 74
Credit: 337135
RAC: 0

RE: RE: Do you recomend

Message 25984 in response to message 25983

Quote:
Quote:
Do you recomend use C37 or not?

This is an unofficial application, so "use at your own risk".

C37 uses FPU (maximum precision) for the calculation, but its needs about 4 bit longer exponent than original.
I checked C37 on my computers with 70-80 units without any error.
A longer test period would be able to determine the frequency of the faults.

I hope that S38 will solve this exponent problem.
It consists a very aggressive precision combination method to extend the precision.

Yeh that's true, I have 33 work units without any error, now I have about 20 units computed by S37a and without any error again. My question was on your (personal) recomandation.
I would like to try S38 but I can't find the link to this version.

On last but not least: thanks again for your graet work :)

Nuadormrac
Nuadormrac
Joined: 9 Feb 05
Posts: 76
Credit: 219177622
RAC: 166750

So far, all my WUs have

So far, all my WUs have validated with C37... I'm wondering when S38 will be posted for us to try? Perhaps to the d/l thread? Again, thx for the binaries :)

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.