PFGW 4.0.4 (with gwnum v30.10) Released

[kar_bon]

Originally Posted by henryzz

I believe you can also do it using the "start" command.

True for Win7/Vista but not for XP.

[rogue]

Originally Posted by Cruelty

Is there a way to set affinity for each instance of pfgw64 / win64pfgw running under windows (other than using task manager)?

A question I have is how much do you really expect to gain by forcing pfgw to run on a specific CPU of a multi-CPU system?

[Antonio]

Rouge,
Out of interest I was trying the -j command line option, however after the run it doesn't appear to produce any output to screen or file!
I have tried both 64 bit MAC and 32 bit WIN versions with the same result (as shown below)

my-macbook:pfgw3562 Toni$ ./pfgw64 -j44497
PFGW Version 3.5.6.64BIT.20111017.Mac_Dev [GWNUM 26.6]

                                              
my-macbook:pfgw3562 Toni$

[rogue]

Originally Posted by Antonio

Rouge,
Out of interest I was trying the -j command line option, however after the run it doesn't appear to produce any output to screen or file!
I have tried both 64 bit MAC and 32 bit WIN versions with the same result (as shown below)

my-macbook:pfgw3562 Toni$ ./pfgw64 -j44497
PFGW Version 3.5.6.64BIT.20111017.Mac_Dev [GWNUM 26.6]

                                              
my-macbook:pfgw3562 Toni$

It is doing the test, but not providing any output for the test result. I'll look into it.

(edit) Upon further investigation, the option is obsolete. The code supports it, but it executes the same code as if specifying -q2^44497-1. I will remove it in the next release. Note that this went away due to changes in gwnum that allows pfgw to use special modular reduction for the form.

[Antonio]

Originally Posted by rogue

It is doing the test, but not providing any output for the test result. I'll look into it.

(edit) Upon further investigation, the option is obsolete. The code supports it, but it executes the same code as if specifying -q2^44497-1. I will remove it in the next release. Note that this went away due to changes in gwnum that allows pfgw to use special modular reduction for the form.

OK, not a problem. I was just attracted by the speed up mentioned in the documentation.

[rogue]

Originally Posted by Antonio

OK, not a problem. I was just attracted by the speed up mentioned in the documentation.

That speed up might have been possible in pfgw 1.x, but changes to gwnum have leveled the field.

[rogue]

Changes from 3.5.5 (3.5.6 was unofficial):

1) Added prevprev() function
2) Fixed ABC2 file with primes from option
3) Fixed issue when providing an argument with -f
4) Removed -j option

You can d/l from here.

[pepi37]

Tell me , does option you write in tutorial is still "active" in latest version ?
Also what is "s" build available in Linux package?



(Note this section will NOT exist by default)
[Woltman_FFTs]
CpuOverride=1 ; 1 or 0. If 1, then CpuType & CpuSpeed are used
CPUtype=?? ; see comment above
CpuSpeed=400 ; The CPU speed of the PC
CpuSupportsRDTSC=1 ; 1 forces usage of RDTSC code, 0 forces not to use
CpuSupportsCMOV=1 ; 1 forces usage of CMOV code, 0 forces not to use
CpuSupportsPrefetch=1 ; 1 forces usage of Prefetch code, 0 forces not to use
CpuSupportsSSE=0 ; 1 forces usage of SSE code, 0 forces not to use
CpuSupportsSSE2=0 ; 1 forces usage of SSE2 code, 0 forces not to use
CpuSupportsMMX=1 ; 1 forces usage of MMX code, 0 forces not to use

Can you clarify what each option do?
Thanks

[rogue]

Originally Posted by pepi37

Tell me , does option you write in tutorial is still "active" in latest version ?
Also what is "s" build available in Linux package?



(Note this section will NOT exist by default)
[Woltman_FFTs]
CpuOverride=1 ; 1 or 0. If 1, then CpuType & CpuSpeed are used
CPUtype=?? ; see comment above
CpuSpeed=400 ; The CPU speed of the PC
CpuSupportsRDTSC=1 ; 1 forces usage of RDTSC code, 0 forces not to use
CpuSupportsCMOV=1 ; 1 forces usage of CMOV code, 0 forces not to use
CpuSupportsPrefetch=1 ; 1 forces usage of Prefetch code, 0 forces not to use
CpuSupportsSSE=0 ; 1 forces usage of SSE code, 0 forces not to use
CpuSupportsSSE2=0 ; 1 forces usage of SSE2 code, 0 forces not to use
CpuSupportsMMX=1 ; 1 forces usage of MMX code, 0 forces not to use

Can you clarify what each option do?
Thanks

You cannot override any of these settings, i.e pfgw writes this data, but does not read it. They are output by gwnum. In fact, I thought that I removed then from the ini file a while ago.

The "s" for Linux is statically linked, as opposed to dynamically linked.

[Prime95]

Originally Posted by rogue

They are output by gwnum.

Um, gwnum does not output this.

[rogue]

Originally Posted by Prime95

Um, gwnum does not output this.

Sorry, I miswrote. I thought that it had. Nevertheless, pfgw has not written this info to the ini file in many releases.