Yeah thanks to all the coders here who are actively porting Alex's V8 code to Windows, Possibly parity at last with the Macs, At least for those that aren't or can't be overclocked, for those that can overclock well I don't need to say that production shouldn't increase any, It should increase a lot, Exactly how much I don't know, But It looks like It will go up none the less. :D :)

---

The T1 Trust, PRR T1 Class 4-4-4-4 #5550, 1 of America's First HST's

Yeah thanks to all the coders here who are actively porting Alex's V8 code to Windows, Possibly parity at last with the Macs, At least for those that aren't or can't be overclocked, for those that can overclock well I don't need to say that production shouldn't increase any, It should increase a lot, Exactly how much I don't know, But It looks like It will go up none the less. :D :)

You are right there Joker. Production will go up when the port is released to the wild.

---

It's good to be back amongst friends and colleagues

Yeah thanks to all the coders here who are actively porting Alex's V8 code to Windows, Possibly parity at last with the Macs, At least for those that aren't or can't be overclocked, for those that can overclock well I don't need to say that production shouldn't increase any, It should increase a lot, Exactly how much I don't know, But It looks like It will go up none the less. :D :)

You are right there Joker. Production will go up when the port is released to the wild.

Thanks, Extra text moved to It's own thread by Your's truly, The Joker.

---

The T1 Trust, PRR T1 Class 4-4-4-4 #5550, 1 of America's First HST's

Yeah thanks to all the coders here who are actively porting Alex's V8 code to Windows, Possibly parity at last with the Macs, At least for those that aren't or can't be overclocked, for those that can overclock well I don't need to say that production shouldn't increase any, It should increase a lot, Exactly how much I don't know, But It looks like It will go up none the less. :D :)

You are right there Joker. Production will go up when the port is released to the wild.

A very good job, J.D.Whale, this surely will work fine on Quad's, f.i. Q6600 .

One off my 2 Q6600's has a clock off 3325MHz, FSB 1486.The other 'stock', with Crunch3r's SSSE3, but the OC'ed beats the stock, running optimized app.

Your result's are looking promising keep up the good works.

---

One off my 2 Q6600's has a clock off 3325MHz, FSB 1486.The other 'stock', with Crunch3r's SSSE3, but the OC'ed beats the stock, running optimized app.

My 2 Q6600's are now running at 3105 (1380 FSB) and 3150 (1400 FSB). It's amasing how well the Q6600's will overclock without really trying.

---

Thanks to all for the work!!

I don't have so much time to read every day this long thread..

I hope the final release of the new apps will be publish in the New Optimized Apps Links -- READ ONLY THREAD that all can profit from it..

When the new app is ready for everybody's use, I am sure there will be an announcement and links posted there.....

---

"Freedom is just Chaos, with better lighting." Alan Dean Foster

Thanks to all for the work!!

I don't have so much time to read every day this long thread..

I hope the final release of the new apps will be publish in the New Optimized Apps Links -- READ ONLY THREAD that all can profit from it..

When the new app is ready for everybody's use, I am sure there will be an announcement and links posted there.....

Sounds good to Me.

---

The T1 Trust, PRR T1 Class 4-4-4-4 #5550, 1 of America's First HST's

JDWale: I'm sure Ozzy would give some devil horns for your effort! Rock on! :-)

---

Host = Thurston: Q6600 01 Apr 02:22 UDT

Update on Thurstons RAC 6500 Challenge shows that RAC climbed ~230 yesterday and it seems the slope trended higher. I haven't even performed any creative scheduling yet, well not very creative anyway ;-) He was even crunching on 3 cores for part of the day while running some benchmarks. So I wonder what's up...?

Any guesses on what day he'll reach 6500 ?
Will he even reach 6500 ?
Any predictions on where he'll top out relative to Marks "Frozen Penny" ?
Maybe BoincStats can help with the predictions, I think I'll go have a look.

[edit] Boincstats says he did 6,432 on 31 Mar... two weeks like that and 6500 is in reach. [/edit]

Cheers,
JDWhale

@OzzFan - Back @ you! Seen Iron Man yet?

---

Host = Thurston: Q6600 01 Apr 02:22 UDT

Update on Thurstons RAC 6500 Challenge shows that RAC climbed ~230 yesterday and it seems the slope trended higher. I haven't even performed any creative scheduling yet, well not very creative anyway ;-) He was even crunching on 3 cores for part of the day while running some benchmarks. So I wonder what's up...?

Any guesses on what day he'll reach 6500 ?
Will he even reach 6500 ?
Any predictions on where he'll top out relative to Marks "Frozen Penny" ?
Maybe BoincStats can help with the predictions, I think I'll go have a look.

[edit] Boincstats says he did 6,432 on 31 Mar... two weeks like that and 6500 is in reach. [/edit]

Cheers,
JDWhale

@OzzFan - Back @ you! Seen Iron Man yet?

Good luck with Thurston......
Just keep in mind that I drop my Frozen Penny back to 2.4v code in between the AK ported test runs I have been doing.....I will not be running the ported code full time until official release.....but the testing sure is fun....LOL.

---

"Freedom is just Chaos, with better lighting." Alan Dean Foster

Good luck with Thurston......
Just keep in mind that I drop my Frozen Penny back to 2.4v code in between the AK ported test runs I have been doing.....I will not be running the ported code full time until official release.....but the testing sure is fun....LOL.

I'm counting on you running 2.4V as the only way Thurston could even come close to catching Penny. Remember that Thurston is only a 3.2GHz Q6600 chasing your 4.58GHz X9650 Extreme... You've got almost 1.5 times the clock rate and 1.5 times the L2 cache. I estimate that my WhalePort of AK code is ~1.6 times the speed of the 2.4V code... This might prove an interresting challenge ;-)

I estimate that if Penny were running the ported code full time that she just might reach RAC of 10,000. I wouldn't stand a chance in ehll of once the "Chicken Plus" port comes public... So a little friendly "exhibition match" to demonstrate ported potential for us "little Q6600 guys" in the mean time.... I might be substituting in other "minor tweaks" to help gain advantage if I can find some, so watch your back, Penny ;-)

Cheers,
JDWhale

---

I'm counting on you running 2.4V as the only way Thurston could even come close to catching Penny. Remember that Thurston is only a 3.2GHz Q6600 chasing your 4.58GHz X9650 Extreme... You've got almost 1.5 times the clock rate and 1.5 times the L2 cache. I estimate that my WhalePort of AK code is ~1.6 times the speed of the 2.4V code... This might prove an interresting challenge ;-)

I estimate that if Penny were running the ported code full time that she just might reach RAC of 10,000. I wouldn't stand a chance in ehll of once the "Chicken Plus" port comes public... So a little friendly "exhibition match" to demonstrate ported potential for us "little Q6600 guys" in the mean time.... I might be substituting in other "minor tweaks" to help gain advantage if I can find some, so watch your back, Penny ;-)

Cheers,
JDWhale

Looks like you might have a chance as long as you don't snag too many VLAR WU's (not much chance!).

According to my records, Thurston is 38% slower than the Frozen Penny at VLAR but 1% FASTER at 0.22 - 0.4 AR and 9% FASTER at 0.4 - 0.5 AR (mean values).

Woo-hoo... Friendly competition.

F.

---

I estimate that if Penny were running the ported code full time that she just might reach RAC of 10,000. I wouldn't stand a chance in ehll of once the "Chicken Plus" port comes public... So a little friendly "exhibition match" to demonstrate ported potential for us "little Q6600 guys" in the mean time.... I might be substituting in other "minor tweaks" to help gain advantage if I can find some, so watch your back, Penny ;-)

Cheers,
JDWhale

Looks like you might have a chance as long as you don't snag too many VLAR WU's (not much chance!).

According to my records, Thurston is 38% slower than the Frozen Penny at VLAR but 1% FASTER at 0.22 - 0.4 AR and 9% FASTER at 0.4 - 0.5 AR (mean values).

Woo-hoo... Friendly competition.

F.

Fred W

I am running a non-frozen Penny on air @ 4.0GHz. This rig uses Win XP Pro and the Crunch3r V2.4 optimised client.

How does a slightly slower Penny, on air, compare to Thurston?

Salivating to the public release of the JD/Alex, Chicken +, code port!

---

It's good to be back amongst friends and colleagues

Fred W

I am running a non-frozen Penny on air @ 4.0GHz. This rig uses Win XP Pro and the Crunch3r V2.4 optimised client.

How does a slightly slower Penny, on air, compare to Thurston?

Salivating to the public release of the JD/Alex, Chicken +, code port!

Though Fred can provide more widespread analysis and total assessment... A quick peak at the 0.39 AR 72.23 credit WUs that we're both choking on right now shows me that Thurston ~16% faster than your Penny at this AR. Of course you're running at 4000Mhz and Thurston is only at 3200Mhz, so the comparison is not at all fair ;-)

Whereas Marks "Frozen" Penny running at 4580MHz with 2.4V SSE4 shows virtually the same CPU times as Thurston running AK WhalePort v0.2 at the same AR.

BOINC On..On...
JDWhale

---

Though Fred can provide more widespread analysis and total assessment... A quick peak at the 0.39 AR 72.23 credit WUs that we're both choking on right now shows me that Thurston ~16% faster than your Penny at this AR. Of course you're running at 4000Mhz and Thurston is only at 3200Mhz, so the comparison is not at all fair ;-)

Whereas Marks "Frozen" Penny running at 4580MHz with 2.4V SSE4 shows virtually the same CPU times as Thurston running AK WhalePort v0.2 at the same AR.

BOINC On..On...
JDWhale

Taking a fairly surface view of your statement ... this suggests, at an AR of ~0.39, that the software of the Chicken+ code is worth at least ~600MHz of CPU clock and the difference in die fabrication (45nM compared to 65nM)? All this on already optimised code!!

---

It's good to be back amongst friends and colleagues

Fred W

I am running a non-frozen Penny on air @ 4.0GHz. This rig uses Win XP Pro and the Crunch3r V2.4 optimised client.

How does a slightly slower Penny, on air, compare to Thurston?

Salivating to the public release of the JD/Alex, Chicken +, code port!

Demonstrating the danger of small sample sizes, your baby looks slightly faster than Thurston at VLAR... I think the other AR's are more representative:


Direct Link

F.

---

Though Fred can provide more widespread analysis and total assessment... A quick peak at the 0.39 AR 72.23 credit WUs that we're both choking on right now shows me that Thurston ~16% faster than your Penny at this AR. Of course you're running at 4000Mhz and Thurston is only at 3200Mhz, so the comparison is not at all fair ;-)

Whereas Marks "Frozen" Penny running at 4580MHz with 2.4V SSE4 shows virtually the same CPU times as Thurston running AK WhalePort v0.2 at the same AR.

BOINC On..On...
JDWhale

Taking a fairly surface view of your statement ... this suggests, at an AR of ~0.39, that the software of the Chicken+ code is worth at least ~600MHz of CPU clock and the difference in die fabrication (45nM compared to 65nM)? All this on already optimised code!!

Mind you that neither Mark nor I are running Chicken+ (sorry if this is not the preferred moniker). Mark might run it for short bursts for benchmarking. I'm running WhalePort v0.2 and am holding nothing back and disclosing everything I've learned and done with the crew working on the Chicken+ port. As said before, I'm restricted by Intel from sharing WhalePort, but I do not want my short-term effort to be lost. I hate to reitterate that point, please understand my position.

By "already optimized", I assume you're referring to the earlier Lunatics KWSN, Cruncher built 2.4V version. While it's true, that code was/is "optimized" and demonstrates a remarkable performance over the "stock" app. The source code in "Alex Kan src-v8.tar" is a total rewrite of several functions that highly leverage the SIMD instruction sets on more than one level (SSE2 SSE3 SSSE3), addin differenly coded algorithms based on processor characteristics and you can see that the code is complex and lends itself to many possible permutations. Give those folks doing all the work at Chicken+ the credit they deserve, That is where your "ported" app will likely come from. I'm just a big mouthed fish that got lucky with some "trial development tools".

IMO...At first & second look, some functions bear absolutely no resembelance to each other. The methods of optimization are beyond my comprehension, I have no working knowledge of SIMD instructions and often find myself laboring over just a few lines of code trying to understand the method. This isn't the stuff you're likely to learn at school without at least a half dozen prerequisites, more likely only after already in the workplace a while and paying Intel dearly for a customized course/workshop. Just my opinion, YMMV.

I don't pretend to be a coding expert and fall far short on Windows code (my working knowledge is in low level 2D Graphics, 3D-Graphics(OpenGL) and performance computing on the SGI Shared Memory MultiProcessor super computers of yesteryear. Like I've stated before, "Kudos to the folks that understand & develop this stuff". I'm just a porter that prefers hitting golf balls.

I think I'll head to the practice range now.

Cheers,
John

---

WU : 26ja07ab.16641.18068.9.7.202
WU true angle range is : 1.242116
CPU time : 3210.625 sec
CPU : type GenuineIntel Intel(R) Pentium(R) Dual CPU E2160 @ 1.80GHz [x86 Family 6 Model 15 Stepping 13]
Number of CPUs : 2
Operating System : Microsoft Windows Vista Home Edition, Service Pack 1,(06.00.6001.00)

----------------------------------------

no comment ? mini 1.27 Angle Range ?

---

SETI@Home Informational message -9 result_overflow
with a general handicap of 80% and it makes much d' efforts for the community and s' expimer, thank you d' to be understanding.

WU : 26ja07ab.16641.18068.9.7.202
WU true angle range is : 1.242116
CPU time : 3210.625 sec
CPU : type GenuineIntel Intel(R) Pentium(R) Dual CPU E2160 @ 1.80GHz [x86 Family 6 Model 15 Stepping 13]
Number of CPUs : 2
Operating System : Microsoft Windows Vista Home Edition, Service Pack 1,(06.00.6001.00)

----------------------------------------

no comment ? mini 1.27 Angle Range ?

WU : 28dc06ag.14133.23794.16.7.218
WU true angle range is : 2.387831
CPU time : 725.108248 sec
CPU type : GenuineIntel Intel(R) Core(TM)2 Quad CPU Q6600 @ 2.40GHz [x86 Family 6 Model 15 Stepping 11]
Number of CPUs : 4
Operating System : Microsoft Windows Vista Professional Edition, (06.00.6000.00)

[edit] wuid=244705636 [/edit]

---

Mind you that neither Mark nor I are running Chicken+ (sorry if this is not the preferred moniker). Mark might run it for short bursts for benchmarking. I'm running WhalePort v0.2 and am holding nothing back and disclosing everything I've learned and done with the crew working on the Chicken+ port. As said before, I'm restricted by Intel from sharing WhalePort, but I do not want my short-term effort to be lost. I hate to reitterate that point, please understand my position.

By "already optimized", I assume you're referring to the earlier Lunatics KWSN, Cruncher built 2.4V version. While it's true, that code was/is "optimized" and demonstrates a remarkable performance over the "stock" app. The source code in "Alex Kan src-v8.tar" is a total rewrite of several functions that highly leverage the SIMD instruction sets on more than one level (SSE2 SSE3 SSSE3), addin differenly coded algorithms based on processor characteristics and you can see that the code is complex and lends itself to many possible permutations. Give those folks doing all the work at Chicken+ the credit they deserve, That is where your "ported" app will likely come from. I'm just a big mouthed fish that got lucky with some "trial development tools".

IMO...At first & second look, some functions bear absolutely no resembelance to each other. The methods of optimization are beyond my comprehension, I have no working knowledge of SIMD instructions and often find myself laboring over just a few lines of code trying to understand the method. This isn't the stuff you're likely to learn at school without at least a half dozen prerequisites, more likely only after already in the workplace a while and paying Intel dearly for a customized course/workshop. Just my opinion, YMMV.

I don't pretend to be a coding expert and fall far short on Windows code (my working knowledge is in low level 2D Graphics, 3D-Graphics(OpenGL) and performance computing on the SGI Shared Memory MultiProcessor super computers of yesteryear. Like I've stated before, "Kudos to the folks that understand & develop this stuff". I'm just a porter that prefers hitting golf balls.

I think I'll head to the practice range now.

Cheers,
John

John, your being too modest.

The AK source has been around for a year or more just waiting for someone to pick up the challenge and port it into a windows application. If it were easy it would have been done long ago by someone else. Your being lucky has nothing to do with it. You accepted the challenge and brought new insight to the problem many had been looking at. It brings truth to the statement that “sometimes a new brain will see something that others have overlooked”.

John, accept your well earned accolades. You have earned your place here.