when under the same task and condition, how faster quad-core will win over dual core?

[coolaler]

when under the same task and condition, how faster quad-core will win over dual core?
platform1
M/B ：ASUS P5B Deluxe
MEM ：G.SKILL 2GB Dual Channel
HDD ：Seagate 250g SATAII 16Mb Cache
VGA CARD ：winfast 6200TC
POWER: Zippy 600WG1
OS ：WIN SEVER 2003
CPU COOLER：Tower 120
CPU ：Intel Core 2 Duo Conroe E6600 S6 4mb cache

platform2
M/B ：ASUS P5B Deluxe
MEM ：ChannelG.SKILL 2GB + XTREEM 2GB Dual Dual Channel
HDD ：Seagate 250g SATAII 16Mb Cache
VGA CARD ：6600GT
POWER: Zippy 850WG1
OS ：WIN SEVER 2003
CPU COOLER：Tower 120
CPU ：Intel Core 2 Quad QX6700 S7 ES 8mb cache

test info
400*9=3.6g
DDRII800 3-3-3-4-5-7-10-5-9

1.Dual spi1M
C2D@3.6Ghz



C2Q@3.6Ghz



See even though the two processors are under same speed, however, quad-core
still got space todeal with super pi 1m, that's why it is faster than dual core 2 seconds.

2.Dual spi32M
C2D@3.6Ghz



C2Q@3.6Ghz



no need to say about 32m, it even fastern than dual core about 30 seconds.

3.Cinebench9.5
C2D@3.6Ghz



C2Q@3.6Ghz



here we can check that out with 3D calculation mode under single core, they both
scaled the same results; but if we use multi-cores to do that, apparently, quad-core would faster than dual-core about 100%.

How about Sandra some benchmarks?
CPU Test1 ; C2D @3.6ghz:



C2Q@3.6Ghz



CPU Test1 ,the score from quad core is 100% off dual core

CPU Test 2
C2D@3.6Ghz



C2Q@3.6Ghz




CPU Test 2,Quad-Core is still about 100% off dual core

how about Memory Bandwidth?
C2D@3.6Ghz



C2Q@3.6Ghz



Dual Core 2DIMM seems larger than quad-core 4DIMM about 100MB/S

L2 Cache

C2D@3.6Ghz



C2Q@3.6Ghz



L2 Cache quad-core 8mb cache under the additional 63% of bandwidth, the speed
is also 100% off dual core.


5.File compression

C2D@3.6Ghz





C2Q@3.6Ghz





5.File Compression

C2D@3.6Ghz





C2Q@3.6Ghz





WinRAR seems like doesn't support multi-tasking, but under the same condition for compressing
32mb folder, quad-core still won over about 7 seconds.

6.3DMARK06 CPU TEST

C2D@3.6Ghz



C2Q@3.6Ghz



C2D 3000pts C2Q 5271pts

7.Encode

C2D@3.6Ghz



C2Q@3.6Ghz



TMPEG supports Multi-Tasking, for the same work of converting 2 files, Quad-Core certainly won over dual core without any problem.

8.SUPERPIRME

C2D@3.6Ghz



C2Q@3.6Ghz



wow!33s VS 17s

9.PCMARK05

C2D@3.6Ghz



C2Q@3.6Ghz



C2Q 11445pts
C2D 9082pts

[HamidFULL]

Great!

[kemo]

great test coolaler that is something I was looking for long time ago
________
starcraft replays

[day187]

thanks cooler

[coolaler]

thanks cooler

what's ur cooler?dude

[Paapaa]

It would be nice to see more actual programs tested and less synthetic benchmarks which rarely reflect the real world difference. Nobody uses (I guess) his computer to calculate millions of digits of pi. So, game tests, office application tests, 3d rendering, OS bootup times etc.

[DAK1640]

Excellent comparisons coolaler, thanks many of us were dying to know. Some people aren't happy unless every known test, bench, game is eval'd...

[wtkprol]

Wow, really nice. I think everybody appreciates the time and effort you put into these tests, so we can see the results. Thanks.

[JMKS]

Great work!
I think that the quad-core is only beneficient (in SPI; in other, multi-threaded, aplications it is obvious that it is a lot faster), because SPI is running on 1st and 3rd core, so the cache is 2x4MB vs 2x[4MB/2] in C2D. There is no other reason, am I right?

[ic3m4n2005]

Have you tried 1M with 2 GB Ram in both machines ? Maybe 4 GB gets an advantage through Interleaving....just a wild guess

[Paapaa]

Some people aren't happy unless every known test, bench, game is eval'd...

No need to show every know test or game. The point was that syntetic benchmarks usually give totally irrelevant information. Those benchmarks don't reflect real application performance differences where the total time is affected by CPU+GPU+IO+Ram.

Let me put it this way: there is no point spending $1000 if it only improves synthetic benchmark scores but real applications/games are not CPU limited at all.

Of course I appreciate the time and effort of Coolaler

[coolaler]

Have you tried 1M with 2 GB Ram in both machines ? Maybe 4 GB gets an advantage through Interleaving....just a wild guess

4GB ram no faster than 2GB ram

[JoeBar]

Thanks for the results coolaler!

[lawrywild]

Nice tests, thanks for your time Coolaler

[BlaqMale]

i appreciate the info sir coolaler

[Pyr0]

excellent, thanks coolaler

just wondering, would or could the shared memory of the TurboCache on the 6200 gfx card make any difference to any results?
is there any chance of testing with 2gb in the quad system please?

[gigabyte137]

Can you run this benchmark on both the c2d 3.6 and the c2q 3.6 and let me know what you get... Thanks... It is a Fritz 9 chess benchmark...

http://www.jens.tauchclub-krems.at/d...0Benchmark.zip

[camouflage]

Wow - one of the best tests I've ever seen - nice work. I vote for

S-T-I-C-K-Y !!!

BTW the point of all is You need at least a fine OCing quad core to benefit from, if it is a low clocker results of some nice running dual cores may be better.

But lets hope the retail quads will be the same insane chips as the duals are...

[erwinz]

Excellent work Coolaler!!

[Rovtar]

nice work

[vmsein]

Thanks coolaler for running the comparisons. This will most certainly help those interested in going quad core decide if the performance is really worth it.

[Cranox]

Nice testing coolaler.

[coolaler]

Can you run this benchmark on both the c2d 3.6 and the c2q 3.6 and let me know what you get... Thanks... It is a Fritz 9 chess benchmark...

http://www.jens.tauchclub-krems.at/d...0Benchmark.zip

ok
10.chess benchmark
C2D@3.6Ghz



C2Q@3.6Ghz

[ochw]

coolaler great work mate keep them coming

[awdrifter]

Very nice. I wish I can encode 2 eps of Shana in 8 mins.