Super Pi 1m scores

[Tatty_Two]

It's not only core 1 but some times it's reversed aswell with very little of core 0 used

That is only because when your puter is on there will always be 2nd core activity no matter how many background proggies you have disabled, yes sometimes Windows (as opposed the th chip) will spread a load across 2 cores (if it can), I did not say your puter wont use 2 cores, I said the benchmark is a single core benchmark, if you could get a single core AMD that had 512KB L2 Cache that ran at 3.5Gig like your 6000 and ran the test in your system with everything else the same the scores would be very similar if you get my meaning.

And again you are confusing load with potential.....don't, whilst they are relative the load on a core is used to requirement, you have to remember there are millions of processes that can be functioning within a core at the same time, depending on what the core is doing will depend on what processes need to be used, for example, if you ran a test or app, if that test or app did not require integer or floating point calculations then that function would not take place within the core process and therefore the load MAY not be full but that does not mean to say that it's not working to it's full potential, a core is about efficiency, not doing what it does not need to do to get a job done if that makes sense? That is actually a very crap and basic way to explain it and it's far too simple really as a CPU is a very complex bit of kit.

 

[trt740]

but look a science mark 2.0 it's uses this chips full potential of both cores and appears to be a better real time test. I would say that super pI only simulates the computers real world power where scientific marks 2.0 actually uses it and demonstrates it more accurately look for yourself. It also tests the memory and how the computer would manage memory and bandwith and appears to be a better rounded test.

 

[Tatty_Two]

No.....your not getting it.....sceincemark 2 is a suite of tests that test most of the CPU's main functions/calculations....that is why the cores are more fully loaded, SuperPI does not.

Edit: And molecular dynamics is a monster of a bench pushing, integer, floating point, code/decode to the limit hence 100%, SuperPI is a completely different kettle of fish.

 

[trt740]

No I get it but what i'm saying is science marks 2.0 shows load and really stress and uses the chips potential. Where super pi mainly simulates it and that science marks 2.0 appear to be more of realistic test. It's not a simulation it's a true test. Simulate science marks and don't run it the right way and your project scores will go up. The real test of a chip is a real world test not a simulation of what it can do. I want to see it acutally do it. How can Super PI be a true test of chips raw power when it doesn't use the chips full core potential and power? I know science marks is because I can see it doing it.

 

[Tatty_Two]

No I get it but what i'm saying is science marks 2.0 shows load and really strees and uses the chips potential. Where super pi manly simulates it and that science marks 2.0 appear to be more of realistic test.

Sceincemark is a system application and performance test, it benches using and testing CPU calculations and process and its communication of that information within the system hence why memory is so important as well as to some degree HDD.

SuperPI is a raw straight number crunching all out, revvved up, Ferrari, mazarati single stream speed test, it does not test all of a CPU's processes like SM2. It's not there to strain the CPU like SM2....just to measure it, make sense? Memory does play some part in it of course as the only way it can manage the info it gets is to read/write to memory.

Phew I am knackered now. But yes, you are right in that respect, it does not test the CPU's FULL potential like SM2.

 

[Tatty_Two]

On an aside, I just bought another CPU!. I just noticed, my post average does not seem to have changed at all over the last few days, its stuck at 4.11 per day no matter how much I post?????

 

[trt740]

okay then why doesn't it use the full power of the cores if it's a true test of the chips power? Also memory management is very important to a chips performance. A chip that doesn't manage memory well would be crippled. So once again I say science mark 2.0 is a better real world whole chip test. okay enough of this talk Tatty what chip did you get?

 

[Tatty_Two]

okay then why doesn't it use the full power of the cores if it's a true test of the chips power? Also memory management is very important to a chips performance. A chip that doesn't manage memory well would be crippled. so once again I say science mark 2.0 is a better real world whole chip test.

Correct, Not power.....speed, they can be different as, as I said, efficiency plays its part.

Edit: Some unknowing person offerewd me the same as I paid for the 4300 and since then the 4400 has been introduced and is now the same price so once it arrives it will be a straight swap.

 

[trt740]

Good deal thats a great chip. Heading out to play ball with my son post some benches when you get that chip.

 

[mk_ln]

14.547

 

[Buckaroo Banzai]

29.641s

 

[ace80]

Still tweaking
13.250s

 

[p_o_s_pc]

I wish i could OC my CPU more but i can't yet I have a new HSF on the way (make it fast ups.)I will post my scores after i get my new HSF and my RAM fan so i can see how high i can get my RAM and CPU maybe i can get a faster time.

 

[caveman_1990]

Is people still updating the result on the first page ?? or are we just sending in pictures for no reason.

 

[giorgos th.]

i am updating the first page..but not know cause i dont have free time..
keep sending results..

 

[Ketxxx]

call me crazy, but once my C2D stuff arrives an I get to size it all up, I'll be aiming for <13s. Whats the fastest 1MB time?

 

[Tatty_Two]

call me crazy, but once my C2D stuff arrives an I get to size it all up, I'll be aiming for <13s. Whats the fastest 1MB time?

Around 10 secs, I think thats at about 3.9Gig on an E6600 with water.

 

[DaEnigma]

1. dominick32: 10.906 (Intel E6700 @ 4700) TOP INTEL SCORE

 

[giorgos th.]

the fastest 1mb time?maybe you mean the fastest 1 million time?

 

[infrared]

Hey mk_ln,

Why are you using such loose timings on the ram? You could get sub 14s if you tightened them up a bit. Try setting 4-4-4-12 and see how much your score improves

the fastest 1mb time?maybe you mean the fastest 1 million time?

Fastest 1M time is sub 9s. It's over at XS if you wanna take a look

 

[giorgos th.]

Thanx Infra...i knew that.

 

[infrared]

oops, misread your post. sorry

 

[zCexVe]

Dont know whether u will like the interference but here is mine.

 

[evoic]

New kid on the block.
C2D E6420 @ 3.0Ghz - 1.3825v
(L651B325 Model) Stock cooling.
2x1GB OCZ REAPER 8500's @ DDR2 1100 / 2.15v
EVGA 122-CK-NF67-A1 680I Board

SuperPi 1.5xs @ 17.03, I've had it under 17, but for some reason it won't do it again.
I'm going to start fooling with the speed as soon as the new cooler arrives.
Shooting for 14's on air.

Oooooh. I just checked the front page and I'm officially the 4th fastest E64'er.

 

[Tatty_Two]

Evoic......what timings are your memory set at?