Overclockers UK Forums Click here for more details
Free Shipping for Loyal Forum Members - CLICK HERE

Go Back   Overclockers UK Forums > Hardware > CPUs

Reply
 
Thread Tools
Old 13th Feb 2011, 10:57   #1
Frank_Rizzo
Wise Guy
 
Joined: Mar 2003
Posts: 1,083
Will 1 sec 1 Million Super Pi ever be achieved?

The CPUs I have had over the years have hit 1M SuperPi times of 40s, 30s, 20s and this week just under 8s.

This got me thinking. Will there ever be a 1 second time for 1M SuperPi?

Based on pure clock speed it looks very much unlikely. The 40/30/20 figures were obtained from 2Ghz/3Ghz/3.8Ghz clock speeds. This sub 8s figure is from a 2500K running at 4.8Ghz

It is now getting so much harder to double clock speed and thus halve SuperPi time. In order to get 1s from 8s the time has to be halved, halved, and halved again. This would mean that the 4.8Ghz speed would have to be doubled, doubled, and doubled again = 38.4Ghz

As SuperPi is single thread, and the clock speed is the biggest determining factor then it does seem that 1 second will never be achieved. At best 5 or possibly 4+ seconds will be world record holders forever.

This is the same as the world record for 100 metres. It will get to a point where humans will no longer physically run any faster, and thus the world record may never be beat (unless the times are started to be recored in .00001 seconds)
Frank_Rizzo is offline   View trust for Reply With Quote
Old 13th Feb 2011, 11:02   #2
KelDG
Hitman
 
Joined: Dec 2009
Posts: 737
Clock speed != Pocessing power

So yes, it will happen, no problem
KelDG is offline   View trust for Reply With Quote
Old 13th Feb 2011, 11:05   #3
Frank_Rizzo
Wise Guy
 
Joined: Mar 2003
Posts: 1,083
Uh for SuperPi clock speed is everything. Single thread.

If clock speeds physically hit a limit of 5/6/7Ghz then I can not see how SuperPi could ever break 4/5 seconds.
Frank_Rizzo is offline   View trust for Reply With Quote
Old 13th Feb 2011, 11:15   #4
KelDG
Hitman
 
Joined: Dec 2009
Posts: 737
Oh, i get it, so it will be program restricted.

That sucks, if thats the case how is it useful as a benchmarking program as that is how I have seen it used.

But even single thread the processing power still increases with each jump forward (q6600 vs sandybridge), plus we dont know what architecture is round the corner. So I still think yes.
KelDG is offline   View trust for Reply With Quote
Old 13th Feb 2011, 11:18   #5
mattyg
Mobster
 
mattyg's Avatar
 
Joined: Jun 2007
Posts: 4,737
Why would clock speeds hit a limit?

-"The trouble with quotes on the Internet is that you can never know if they are genuine"
- Abraham Lincoln
mattyg is online now   View trust for Reply With Quote
Old 13th Feb 2011, 11:32   #6
El Jimben
Wise Guy
 
El Jimben's Avatar
 
Joined: Jul 2006
Location: Essex
Posts: 2,196
Quote:
Originally Posted by Frank_Rizzo View Post
Uh for SuperPi clock speed is everything. Single thread.

If clock speeds physically hit a limit of 5/6/7Ghz then I can not see how SuperPi could ever break 4/5 seconds.
Not entirely true as the amount of work done per clock cycle plays a part too, you won't see a 3GHz Pentium D beating a 2.8GHz C2D for instance.

3570K @ 4.4GHz||GTX 970||8GB DDR3||MSI Z68-GD65
El Jimben is offline   View trust for Reply With Quote
Old 13th Feb 2011, 11:57   #7
james.miller
Capodecina
 
james.miller's Avatar
 
Joined: Aug 2003
Location: Woburn Sand Dunes
Posts: 14,605
im not sure how relevant super pi is these days. It's interesting to know y-cruncher can calculate 50 million digits faster than superPI can can calculate 1 million on the same cpu (phenom2-965 at stock)

Quote:
Validation Version: 1.1

Program: y-cruncher - Gamma to the eXtReMe!!! ( www.numberworld.org )
Copyright 2008-2011 Alexander J. Yee ( a-yee@northwestern.edu )


User: None Specified - You can edit this in "Username.txt".


Processor(s): AMD Phenom(tm) II X4 965 Processor
Logical Cores: 4
Physical Memory: 4,293,058,560 bytes ( 4.00 GB )
CPU Frequency: 3,411,630,543 Hz

Program Version: 0.5.5 Build 9178 Alpha (x64 SSE3 - Windows ~ Kasumi)
Constant: Pi
Algorithm: Chudnovsky Formula
Decimal Digits: 100,000,000
Hexadecimal Digits: Disabled
Threading Mode: 4 threads
Computation Mode: Ram Only
Swap Disks: 0
Working Memory: 536 MB

Start Date: Mon Feb 14 11:57:49 2011
End Date: Mon Feb 14 11:58:40 2011

Computation Time: 48.331 seconds
Total Time: 51.429 seconds

CPU Utilization: 377.12 %
Multi-core Efficiency: 94.28 %

Last Digits:
9948682556 3967530560 3352869667 7734610718 4471868529 : 99,999,950
7572203175 2074898161 1683139375 1497058112 0187751592 : 100,000,000

Timer Sanity Check: Passed
Frequency Sanity Check: Passed
ECC Recovered Errors: 0
Checkpoint From: None

----

Checksum: 43bd4fe66fdd58711699abc29c81e3a4773c3132dc621ed00e 026ab76d56cddd

Last edited by james.miller; 13th Feb 2011 at 11:59.
james.miller is offline   View trust for Reply With Quote
Old 13th Feb 2011, 12:00   #8
Frank_Rizzo
Wise Guy
 
Joined: Mar 2003
Posts: 1,083
Cache size important too and this has helped figures over the years.

IMO 1 second will never be achieved with current home use technology. 10GHz+ surely can't be achieved and no new chipset, increase in RAM bandwidth and decrease in latencies will be sufficient enough to achieve 1 second.

Yes there are other ways of calculating pi figues. And SuperPi is now probably as irrelevant as achieving the highest score on Donkey Kong, whilst everyone else is amassing points in some MMO.

But just for the record. I'm sure there will be a world record set for SuperPi of around 4+ seconds and it will never be beat. That's the challenge ...
Last edited by Frank_Rizzo; 13th Feb 2011 at 12:03.
Frank_Rizzo is offline   View trust for Reply With Quote
Old 13th Feb 2011, 12:13   #9
james.miller
Capodecina
 
james.miller's Avatar
 
Joined: Aug 2003
Location: Woburn Sand Dunes
Posts: 14,605
Quote:
Originally Posted by Frank_Rizzo View Post
Yes there are other ways of calculating pi figues. And SuperPi is now probably as irrelevant as achieving the highest score on Donkey Kong, whilst everyone else is amassing points in some MMO.

that wasnt my point. SuperPI was always used a performance benchmark but i dont think it's relevant any more. At least with y-cruncher its about how fast we can calculate Pi. SuperPI is neither the fastest method of calculation, nor is it indicative of modern-day cpu performance. I'm left wondering why we still hold on to it - you dont see people benching 3dmark 2003 anymore, for example.

Last edited by james.miller; 13th Feb 2011 at 12:18.
james.miller is offline   View trust for Reply With Quote
Old 13th Feb 2011, 12:42   #10
mattyg
Mobster
 
mattyg's Avatar
 
Joined: Jun 2007
Posts: 4,737
There will NEVER become a point that Technology stops.

Back in the P3 450 days the magazines were saying that we wont be able to get above 1Ghz. The technology simply doesnt exist.

Well it does now. And the same will happen in the future.

-"The trouble with quotes on the Internet is that you can never know if they are genuine"
- Abraham Lincoln
mattyg is online now   View trust for Reply With Quote
Old 13th Feb 2011, 13:00   #11
Frank_Rizzo
Wise Guy
 
Joined: Mar 2003
Posts: 1,083
So it may be possible to get 38Ghz and thus 1 second in about ... 15-20 years?

I'm still not convinced. The speed barrier seems to be the biggest hold back at the moment. In the 80's and 90's it was easy to double processor speeds from 16Mhz to 33Mz .. 100Mhz, 200Mhz, 400Mhz. It sort of followed Moore's Law.

But in the past 5 or so years clock speed has been harder to increase and thus multi cores, bigger caches are used for performance gains. Thus with current architecture I don't think there will ever be stock speeds of 10Ghz, let alone the 30Ghz+ which would need to be required for 1M SuperPi.

It's like when the first jet engines were trialled. The sound barrier was a restriction. Of course, mach X speeds were achieved later once restrictions were overcome but it was decades later that it was available for commercial use.

1 second may be possible in 20 years but it would be a totally different architecture to todays technology. Maybe silicon has had it's days.
Frank_Rizzo is offline   View trust for Reply With Quote
Old 13th Feb 2011, 13:51   #12
mattyg
Mobster
 
mattyg's Avatar
 
Joined: Jun 2007
Posts: 4,737
Thread bookmarked. So I can come back in 10 yrs and say "I told you so"

-"The trouble with quotes on the Internet is that you can never know if they are genuine"
- Abraham Lincoln
mattyg is online now   View trust for Reply With Quote
Old 13th Feb 2011, 14:00   #13
Frank_Rizzo
Wise Guy
 
Joined: Mar 2003
Posts: 1,083
I was going to offer odds of 100/1 that it would not be achieved within 10 years.

I think there would be too many takers
Frank_Rizzo is offline   View trust for Reply With Quote
Old 13th Feb 2011, 14:01   #14
Mik3
Mobster
 
Mik3's Avatar
 
Joined: Oct 2002
Location: Andromeda
Posts: 2,763
Probably not with silicon, but maybe with some other medium in the future.

Mik3 is offline   View trust for Reply With Quote
Old 13th Feb 2011, 14:04   #15
Frozennova
Soldato
 
Joined: Nov 2009
Location: Northampton
Posts: 6,765
Your forgetting the fact that a 1ghz pentium 3 will run significantly slower than a 1ghz Sandy bridge cpu, its not just clock speed increase that will improve times, memory latencies are architectural improvements also bring performance increases.


Celeron G1610 | 4Gb Mushkin Radioactive | GTX460
Frozennova is offline   View trust for Reply With Quote
Old 13th Feb 2011, 14:54   #16
Thejay
Gangster
 
Joined: Oct 2009
Posts: 262
http://hwbot.org/rankings/benchmark/superpi/rankings

There you go 6s has been broken
Thejay is offline   View trust for Reply With Quote
Old 13th Feb 2011, 16:41   #17
weird_dave
Hitman
 
Joined: Mar 2007
Posts: 886
I expect 5s to be broken by the end of summer and 4s early next year, LN2 ***!
weird_dave is offline   View trust for Reply With Quote
Old 13th Feb 2011, 18:09   #18
gug54321
Gangster
 
Joined: Jan 2011
Posts: 166
just checked mine 19s for stock 1090t whopee lol

Gigabyte 890GPA-U3DH :: Coolermaster Scout :: AMD1090T 6x 3.2Mhz ::ATI HD6870 :: Zalman CNPS9900 :: Cyborg V7 keyboard ::OCZ 650w PSU :: Western dig 1TB + 1.5TB HDD :: GSkill Ripjaws 4x2Gb ::
gug54321 is offline   View trust for Reply With Quote
Old 13th Feb 2011, 19:12   #19
man_from_uncle
Wise Guy
 
Joined: Jan 2006
Posts: 1,345
It'll get broken in a couple of generations, maybe three, but no more.

As others have mentioned more efficient methods of using each clock have continued to be found. Pentium 3 was actually more efficient than Pentium 4 clock for clock which is why, when Intel found there were physical problems getting processors to run much faster than 4ghz (rather than the 10+Ghz they thought would eventually be achievable using P4), they switched back to framework with a more efficient architecture.

Core 2 took that to a new level, then i3/5/7 beat that.

In 2 Generations I think we will see sub 1 second times when the L2N crowd get hold of those new parts. Then it'll be another generation before we start seeing people do it on Water or Air.

Opty 170 @ 2.7 (10x270) 1.5v
2x1gig XMS3500LL 225@2-3-2-6 2.7v
Asus A8N-SLi Premium, Zalman 9500, 3x80 gig RAID 5
2 X 7900gt SLi @ 550/1600 2 X Zalman vf900
man_from_uncle is offline   View trust for Reply With Quote
Old 13th Feb 2011, 19:18   #20
JonJ678
Capodecina
 
Joined: Dec 2008
Location: England
Posts: 10,246
Quote:
Originally Posted by KelDG View Post
Clock speed != Pocessing power
This is such a simple point that I'm astonished that some people don't get it.

Quote:
Originally Posted by Lysander View Post
Dreams can, occasionally, be predictive since the subconscious mind has the ability to see into the future.
JonJ678 is offline   View trust for Reply With Quote
Old 13th Feb 2011, 20:34   #21
Frank_Rizzo
Wise Guy
 
Joined: Mar 2003
Posts: 1,083
And skimming threads without reading all of the information is such a simple taks that I am astonished some people do it.

SuperPi is single thread. The biggest influence on the performance of the app is the clock speed. The higher the clock speed the better the results. Adding 4, 8, 12 cores, increasing the cache by a Meg or so, increasing memory bandwidth will hardly make any impact on the results.

Look at the results on the hwbot site. There is a very strong correlation between performance and clock speed.
Frank_Rizzo is offline   View trust for Reply With Quote
Old 13th Feb 2011, 20:44   #22
Bramill
Gangster
 
Joined: Sep 2009
Posts: 135
There used to be a joke i think in the sixties that one day you'll be able to buy a computer in a box. I think the future will be the same, in that what we have now is laughable compared to whats possible.

Quantum computers are in development, some people reckon they will be able to break the strongest encryption in use today in seconds rather than tens of thousands of years.

JJ
Last edited by Bramill; 13th Feb 2011 at 20:46. Reason: grammar
Bramill is offline   View trust for Reply With Quote
Old 13th Feb 2011, 20:47   #23
Orcvader
Capodecina
 
Orcvader's Avatar
 
Joined: Oct 2009
Location: Essex, England
Posts: 11,235
Well I'm sure it may be possible one day if this picks up: http://www.engadget.com/2010/02/07/i...ne-transistor/

We're reaching the limits of silcon very quickly now .

Orcvader is offline   View trust for Reply With Quote
Old 13th Feb 2011, 21:16   #24
Gashman
Mobster
 
Joined: Mar 2006
Location: Jarrow, Tyne And Wear
Posts: 4,379
super-pi is pointless, thats pretty much it to be honest. whats the point in todays world of a so called 'benchmark' program being single threaded? it won't happen unless its made to be multi-threaded, since that is the obvious direction things are heading, to use super-pi as an actual benchmark of a CPUs performance isn't relevant anymore, since the number of cores is growing, core size is shrinking. its all heading in the direction of efficiency over outright power, better multi-threaded performance is the 'real' benchmark IMO, anything else is pointless. like trying to clock the speed of a cat with one leg or something, you just don't do it!

Bulldozer FX-8120 4.6GHZ // [COLOR=DarkOrangeGigabyte 990FXA-UD3[/color] // 16Gb Kingston HyperX DDR3 // MSI HD7850 // Samsung Spinpoint 750Gb x 2 // OCZ Z-Series PSU // Custom Watercooling Setup //
Gashman is offline   View trust for Reply With Quote
Old 14th Feb 2011, 00:17   #25
Frank_Rizzo
Wise Guy
 
Joined: Mar 2003
Posts: 1,083
What's the point in continuing to conquer Everest?
What's the point in testing the new Focus over 0-62?
What's the point in continuing the Paris to Dakar Rally?
What's the point in benching rigs with SuperPi?

It's a challenge to be taken on.

My OP point is that I don't think 1 second will ever be achievable. This is much in the same way that no human with current body (chipset) and mindset (processor) will ever break the 3 minute mile, as there has to be a point where current physicality (technology) can no longer achieve a record.

I think that a 4 second 1M SuperPi will be the fastest ever and not ever beaten. If 4s is achieved it will be by silicon on something which would run at around 8-9Ghz, possibly in 2012/13 and that is it.

Technology always improves but it would need a sea-change in structure to be able to break the 1s 1M SuperPi barrier. A pointless exercise maybe. But for users which require pure grunt performance, rather than multiple processes handling all things at once, it is important.
Last edited by Frank_Rizzo; 14th Feb 2011 at 00:28.
Frank_Rizzo is offline   View trust for Reply With Quote
Old 14th Feb 2011, 00:59   #26
KelDG
Hitman
 
Joined: Dec 2009
Posts: 737
Ok, I have one for you. You can run a snes emulator to run a snes game on the PC 100 of times faster than you could on the original hardware.

What if a few generations of processor down the line you emulate a "single core pc" on the however many core wierd new architecture pc of the future. You could emulate a 100gz processor (effectivley letting the multiple cores power the emulation of a single core pc), just the same way you can emulate a 21mhz Snes now.
KelDG is offline   View trust for Reply With Quote
Old 14th Feb 2011, 02:19   #27
Gashman
Mobster
 
Joined: Mar 2006
Location: Jarrow, Tyne And Wear
Posts: 4,379
Quote:
Originally Posted by KelDG View Post
Ok, I have one for you. You can run a snes emulator to run a snes game on the PC 100 of times faster than you could on the original hardware.

What if a few generations of processor down the line you emulate a "single core pc" on the however many core wierd new architecture pc of the future. You could emulate a 100gz processor (effectivley letting the multiple cores power the emulation of a single core pc), just the same way you can emulate a 21mhz Snes now.
in the future, where programs will be more efficiently threaded, etc. why would you want to have a multi-core processor act as a single core processor, surely that would defeat the purpose of progression? also what is the point in running a game a hundred times faster than it originally ran, if it originally ran totally fine in the first place? think about it this way, supreme commander, which is pretty good use of multiple cores will significantly faster (especially in very large scale battles, etc.) on a Phenom II X6 than it would on a 4GHZ C2D, however willing to say the C2D would annihilate the Phenom in Super-PI, so at the end of the day does it really give an accurate representation of performance, and if not where is its validity as a benchmark tool, don't get me wrong in the past it was perfect but i feel technology and the market are moving away from that kind of application.

also Frank_Rizzo i get what your saying, i really do but at the end of the day for me the market is heading in one direction, efficiency and that for the most part means handling tasks more efficiently. would prefer my quad-core to be using as much of its resources as possible as often as possible rather than being 'really' fast in a situation that effectively wastes 75% of its 'potential' performance, but not utilizing the three other cores, so just because a 2.4GHZ Q6600 scores lower in Super-PI than a 3GHZ E6600 for example, does it mean that its less powerful, in the overall scheme of things no, just inefficiently utilized. guess we'll have to just agree to disagree on this one. but if i had to answer, would guess graphene based processors could break the 1s mark due to sheer clock speed, however i don't think any future graphene processor will be geared toward single-threaded applications.

Bulldozer FX-8120 4.6GHZ // [COLOR=DarkOrangeGigabyte 990FXA-UD3[/color] // 16Gb Kingston HyperX DDR3 // MSI HD7850 // Samsung Spinpoint 750Gb x 2 // OCZ Z-Series PSU // Custom Watercooling Setup //
Gashman is offline   View trust for Reply With Quote
Old 14th Feb 2011, 04:10   #28
Psycho Sonny
Capodecina
 
Psycho Sonny's Avatar
 
Joined: Jun 2006
Posts: 20,110
Quote:
Originally Posted by Frank_Rizzo View Post
The CPUs I have had over the years have hit 1M SuperPi times of 40s, 30s, 20s and this week just under 8s.

This got me thinking. Will there ever be a 1 second time for 1M SuperPi?

Based on pure clock speed it looks very much unlikely. The 40/30/20 figures were obtained from 2Ghz/3Ghz/3.8Ghz clock speeds. This sub 8s figure is from a 2500K running at 4.8Ghz

It is now getting so much harder to double clock speed and thus halve SuperPi time. In order to get 1s from 8s the time has to be halved, halved, and halved again. This would mean that the 4.8Ghz speed would have to be doubled, doubled, and doubled again = 38.4Ghz

As SuperPi is single thread, and the clock speed is the biggest determining factor then it does seem that 1 second will never be achieved. At best 5 or possibly 4+ seconds will be world record holders forever.

This is the same as the world record for 100 metres. It will get to a point where humans will no longer physically run any faster, and thus the world record may never be beat (unless the times are started to be recored in .00001 seconds)
records are broken on a daily basis. lets take your 100m sprint for example. as medicine and genetics advance so will steroids and humans , i expect to see the 100m time broken several times within my life time. technology will always advance and so will humans. there are also different types of cpu's , i'd like to see a ps3 or a gpu run this test. it will be doable within 1 second, you have proved this yorself by stating how your times have dropped significantly within a short space of time, it may take 5 years or 10 but it will happen

PSN - PsychoSonny Xbox Live - PsychoSonny
Psycho Sonny is offline   View trust for Reply With Quote
Old 14th Feb 2011, 10:48   #29
KelDG
Hitman
 
Joined: Dec 2009
Posts: 737
Quote:
Originally Posted by Gashman View Post
in the future, where programs will be more efficiently threaded, etc. why would you want to have a multi-core processor act as a single core processor, surely that would defeat the purpose of progression? also what is the point in running a game a hundred times faster than it originally ran, if it originally ran totally fine in the first place? think about it this way, supreme commander, which is pretty good use of multiple cores will significantly faster (especially in very large scale battles, etc.) on a Phenom II X6 than it would on a 4GHZ C2D, however willing to say the C2D would annihilate the Phenom in Super-PI, so at the end of the day does it really give an accurate representation of performance, and if not where is its validity as a benchmark tool, don't get me wrong in the past it was perfect but i feel technology and the market are moving away from that kind of application.
Sorry, you misunderstand me. I don't think there is any point in SuperPI, but this is a discussion on how to get SuperPI running faster than 1 second, it was just an idea on how to do that. Basically a way to run a single thread by using all the cores of a future PC by emulating a single core PC. To be fair this is a long way round sidestepping the fact that it probably would be easier to re-wright SuperPI to use multiple threads, but that isn't the point of this thread.
KelDG is offline   View trust for Reply With Quote
Old 14th Feb 2011, 11:47   #30
man_from_uncle
Wise Guy
 
Joined: Jan 2006
Posts: 1,345
I'm surprised so many people doubt that more efficient CPU design will make up for the finite nature of Silicone.

It's not a matter of cache or threads or cores, it's a matter of how many clock cycles it takes to perform an action. For example Netburst needed (for the sake of arguement) say 16 cycles to perform an calculation. Pentium III needed only 12. AMD64 needed 10. Core2 needed 8. i3/5/7 needs 7 and within a few years the next generation will need come out that will be able to get more work done per clock cycle again.

SuperPi 1M will be done in under a second, and within 3-4 generations it'll be achievable with a CPU at only 3-4 Ghz. That's 10 years from now, but i suspect the hardcore crowd will get there in about half that time.

Opty 170 @ 2.7 (10x270) 1.5v
2x1gig XMS3500LL 225@2-3-2-6 2.7v
Asus A8N-SLi Premium, Zalman 9500, 3x80 gig RAID 5
2 X 7900gt SLi @ 550/1600 2 X Zalman vf900
man_from_uncle is offline   View trust for Reply With Quote
Reply

Thread Tools

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Forum Jump


All times are GMT. The time now is 17:11.


Powered by vBulletin®
Copyright ©2000 - 2016, Jelsoft Enterprises Ltd.
Overclockers UK (Ocuk Ltd)