Benchmark your Computer with the New Geekbench 3

[Abwx]

I don't have the ICC_patch unfortunately . I tried googling but found no reliable source to DL it from.

http://www.softpedia.com/progDownload/Intel-Compiler-Patcher-Download-154486.html

There are two links in the right of the page , one from
a US server and another from a UK one.

 

[BallaTheFeared]

Worked for me in 8.1 preview, at least it let me apply it and found 3 files to patch.

Of course I didn't use it because I wouldn't want to disable my cheat.

 

[Abwx]

Worked for me in 8.1 preview, at least it let me apply it and found 3 files to patch.

Of course I didn't use it because I wouldn't want to disable my cheat.

Let say that granted there s differences it would be more
unequal optimization than cheating , only thing is that some
CPUs wouldnt work at their real potentials but the bench
would still be relevant when comparing intel only CPUs
or even AMDs only if they are of a same generation.

 

[inf64]

Thanks Abwx . I will try it later and see what happens.

 

[FalseChristian]

I know nobody gives a shit but here are my scores:

-Intel Core i5 2500K 3.3GHz @ 4.5GHz (45x100)
-2x4GB Kingston ValuRam DDR3 1333MHz @ 1600MHz

-Single-Core score: 3690
-Multi-Core score : 12137

 

[BallaTheFeared]

I can't wait to see the huge performance gains for AMD once the cripple_amd function is removed!

 

[Nothingness]

And even better considering it s compiled with ICC
with numerous "intel only paths"....

Are you sure this proves it is icc? Isn't it geekbench code for CPU identification? As far as I know VC is used for Windows compilation.

 

[Piklar]

4770K @4.6 before ICC_patch .

4770K @4.6ghz after ICC_patch . Hardly any difference..

 

[Abwx]

Are you sure this proves it is icc? Isn't it geekbench code for CPU identification? As far as I know VC is used for Windows compilation.

As per the devellopper infos:

Intel Compiler Patcher (ICP) scans your hard drive for executable files compiled with the Intel C++ Compiler making it possible to disable the CPU dispatcher in detected files, thus, increasing performance of the software that uses these files with CPUs other than Intel.

The CPU dispatcher function used in Intel compilers is represented by the following piece of code:


[..skip..] mov eax, [ebp][-0008] cmp eax, 0756E6547 ;"uneG" ; Checking on "Genu" jne not_intel ; if it doesn’t equal, switching for not_intel mov eax, [ebp][-0010] cmp eax, 049656E69 ;"Ieni" ; Checking on "ineI" jne not_intel ; doesn’t equal - not_intel mov eax, [ebp][-0014] cmp eax, 06C65746E ;"letn" ; Checking on "ntel" jne not_intel ; doesn’t equal - not_intel mov edx, 000000001 ; the secret byte jmps next not_intel: xor edx, edx ; and here we have 0 for all non-Intel CPUs next: [..skip..]


We can see that the CPU dispatcher checks on "Genu", "ineI", "ntel" and if all checks are successful EDX is given value “1” (one) meaning Intel CPU. Otherwise – “0” (zero) meaning CPU made by any other vendor. Thus, if these checks are disabled or replaced with a condition that is always true in the end EDX register will have ‘1’ (one) regardless of what CPU is actually used.

 

[Nothingness]

As per the devellopper infos:

Dev info from what?

Any CPU identification function would need comparison against that string; without seeing the surrounding code it's hard to say this proves it is icc.

 

[Abwx]

Dev info from what?

Any CPU identification function would need comparison against that string; without seeing the surrounding code it's hard to say this proves it is icc.

The devellopper of the patch....


A path that is choosen after identification of "genuine intel"
cant be the result of MS or GCC CPU dispatcher.

 

[Nothingness]

A path that is choosen after identification of "genuine intel"
cant be the result of MS or GCC CPU dispatcher.

Have you seen identification code used to find CPU model name? Yes it is cmp followed by jumps.

I don't say you're wrong, but given that John Poole said Windows Geekbench is using VC and that he seems to have been reliable, I'd be very surprised to find he is in fact using icc.

And to convince me you'll have to show me full disassembly of the offending code... That is unless I decide to reverse engineer it myself

 

[Abwx]

The linux version use GCC , indeed.
The subject would deserve its own thread anyway ,
you can always download said utility and check by yourself
its relevancy and eventual drawbacks.

 

[Sweepr]

Not bad for a $200 hasfail

Damn, thats nearly twice the single thread score of AMD's Piledriver cores running at similar frequency.

Nehalem's AES subscore is very low (absense of AES extension), thats probably bringing the overall results down on older CPUs.

2C/4T Ivy Bridge ULV i7-3517U @ 1.9-3.0GHz: 2466 Single-core & 5164 Multi-Core.
22.3% slower than a 4C desktop Richland @ 4.3GHz (MT) and 8.3% faster @ single-thread.
Not bad for a 17W CPU+iGPU ultraportable chip.

 

[BallaTheFeared]

Not fair, low blow, until the extreme/water coolers show up that's the #1 single thread score on the site so far.

The second fastest 32 bit score is 4405.

 

[SlowSpyder]

Thuban @ 4.03GHz (DDR2 @ 866MHz, if it matters)
Single core - 2048
Multi core - 10796

 

[Udgnim]

4.6 GHz 4770K w/ HT disabled, DDR3 1333

4134 / 13767

4.6 GHz 4770K w/ HT enabled, DDR3 1333

4132 / 16174

4.6 GHz 4770K w/ HT enabled, DDR3 1600

4216 / 16644

4.7 GHz 4770K w/ HT enabled, DDR3 1600

4285 / 16914

4.8 GHz 4770K w/ HT enabled, DDR3 1600

4367 / 17105

4.9 GHz 4770K w/ HT enabled, DDR3 1600

4448 / 17587

Spoiler

5GHz would not boot into Windows 7 using 1.27V

oh well don't care enough to increase voltages

 

[Engineer]

Single-Core Score Multi-Core Score
3149 / 12126

Intel Xeon E3-1230V2 @ 3.29 GHz

System Information
MSI MS-7759
Operating System Microsoft Windows 7 Professional (64-bit)
Model MSI MS-7759
Processor Intel Xeon E3-1230V2 @ 3.29 GHz
1 processor, 4 cores, 8 threads
Processor ID GenuineIntel Family 6 Model 58 Stepping 9
Processor Codename Ivy Bridge
Processor Package Socket 1155 LGA
L1 Instruction Cache 32 KB x 4
L1 Data Cache 32 KB x 4
L2 Cache 256 KB x 4
L3 Cache 8192 KB
Motherboard MSI Z77MA-G45 (MS-7759)
Northbridge Intel ID0158 09
Southbridge Intel Z77 04
BIOS American Megatrends Inc. V1.7
Memory 32768 MB -1MHz

 

[monstercameron]

http://browser.primatelabs.com/geekbench3/8573

​

Single-Core Score| Multi-Core Score
2063 | 7358
PII x4 965BE @ 4.1Ghz w/ ddr3 1600MHz


run 1:http://browser.primatelabs.com/geekbench3/8613
run 2:http://browser.primatelabs.com/geekbench3/8797
​

Single-Core Score| Multi-Core Score
428/430 | 742/761
amd c-50 @ 1GHz w/ ddr3 1066MHz

notes: tried the android version on my c-50 netbook the results were 360/~6xx(crashed before I could get the second number) which is a regression from the windows results
​

 

[inf64]

I'm happy with my almost nehalem like performance from 4T 80 bucks chip . For what I use it for it's even an overkill lol .

 

[YBS1]

Not fair, low blow, until the extreme/water coolers show up that's the #1 single thread score on the site so far.

The second fastest 32 bit score is 4405.

I have serious doubts anyone who frequents this forum is going to beat your single core score. AFAIK you have the only haswell on here over 4.8GHz correct? Hyperthreading isn't giving enough boost in this benchmark for the 4770K'ers to match you without being probably at or around 4.8.

Edit: Nevermind, saw the 4770K@4.9 above, still not enough. What were you clocked at while running this?

 

[SPBHM]

I'm happy with my almost nehalem like performance from 4T 80 bucks chip . For what I use it for it's even an overkill lol .

nehalem is from 2008.

you should compare it with newer CPUs, like the i3 also from 2012

http://browser.primatelabs.com/geekbench3/compare/8459?baseline=7776

your CPU does extremely well for AES (no surprise, thanks to Intel disabling AES NI for i3), and some of the integer test, but not so amazing on floating point, considering the 1GHz higher clock

still the X4 750K is definitely a great CPU for the money, with your OC faster than a stock 6800K I guess

 

[BallaTheFeared]

I have serious doubts anyone who frequents this forum is going to beat your single core score. AFAIK you have the only haswell on here over 4.8GHz correct? Hyperthreading isn't giving enough boost in this benchmark for the 4770K'ers to match you without being probably at or around 4.8.

Actually that new guy has a better Haswell chip than me, be careful he might take it over 1.27v


i5 @ 5Ghz core 5GHz cpuNB

 

[YBS1]

Actually that new guy has a better Haswell chip than me, be careful he might take it over 1.27v


i5 @ 5Ghz core 5GHz cpuNB

Yeah, I noticed that nice low voltage at 4.9 as well, hyperthreading still enabled even. He just might, and if he doesn't that's just wrong for him to get such a chip. :awe:

 

[monstercameron]

nehalem is from 2008.

you should compare it with newer CPUs, like the i3 also from 2012

http://browser.primatelabs.com/geekbench3/compare/8459?baseline=7776

your CPU does extremely well for AES (no surprise, thanks to Intel disabling AES NI for i3), and some of the integer test, but not so amazing on floating point, considering the 1GHz higher clock

still the X4 750K is definitely a great CPU for the money, with your OC faster than a stock 6800K I guess

my 965BE compares better[still a shame] but memory performance is way low...
http://browser.primatelabs.com/geekbench3/compare/8459?baseline=8573