Benchmark your computer with Handbrake

[JoeRambo]

Video encoding is mostly (entirely?) integer though. My guess its the front end that's the bottleneck as a quad core Ph2 can decode 12 instructions compared a 2 module FX's 8.

Yup. It's mostly integer based, so for example they are using AVX2, but little AVX ( mostly VEX encoded instructions, but no 256b FP SIMD ).

 

[24601]

3570k@4.6ghz@1.296v
4x8gB Crucial Ballistix Sport LP 1.3v@2000mhz 9-9-9-24-2N@1.635v

HandBrake svn5622 (Nightly Build) - 64bit Version
OS: Microsoft Windows NT 6.1.7601 Service Pack 1
CPU: Intel(R) Core(TM) i5-3570K CPU @ 3.40GHz
Ram: 32725 MB, Screen: 1920x1080

x264 [info]: using cpu capabilities: MMX2 SSE2Fast SSSE3 SSE4.2 AVX
[08:11:17] work: average encoding speed for job is 229.081482 fps

Slow.

 

[Uhtrinity]

fx8320 @ 4Ghz, 8GB DDR3


[07:25:13] starting job
[07:26:19] work: average encoding speed for job is 221.016724 fps

 

[Haserath]

There is no bottleneck as FX8xxx just crushes every other AMD chip ( OCed X6 included). You can see in the table that 1st gen BD based FX6100 @ 4.1Ghz scores ~170pts vs ~184pts (average ) of X6 @ 4Ghz. That's basically ~10% difference while having 2x less dedicated FP hardware-an amazing showing by BD/PD .

PS Anything SSE (packed integer) is still done on FlexFP unit.

Just because Bulldozer's unit isn't dedicated to 1 core doesn't mean it's been kept the same.

It has nearly double the resources. If it had kept the same width, it would be half of Phenom II for a dual module.

 

[inf64]

I'm talking about the sheer number of avaliable pipelines. There is no way around it: Bulldozer has 2x less dedicated pipelines but they are multipurpose and FlexFP uses SMT to fill in the bubbles. There are many other improvements(mov optimization for example) that help the FP unit reach this peak efficiency. It's a testament of the design. In the end FX8xxx performs much like i7: 4 dual threaded FP units Vs 4 dual threaded FP units. X6 is just left in the dust and this is with no FMA in mind. Add the FMA3/4 and BD's peak FP exec. throughput is almost doubled.

 

[ElFenix]

HandBrake 0.9.9.5530 - 32bit Version
OS: Microsoft Windows NT 6.0.6002 Service Pack 2
CPU: AMD Athlon(tm) 64 X2 Dual Core Processor 3600+
Ram: 3069 MB, Screen: 1920x1080

x264 [info]: using cpu capabilities: MMX2 SSE2Slow SlowCTZ
[12:18:23] work: average encoding speed for job is 15.452947 fps

@ 1.9GHz

 

[Makaveli]

Pretty impressed with how my old chip is keeping up.

A 3930k at the same 4.2 Ghz only scores 32 points more.

I was expecting abit more with the IPC improvements SB has and access to the AVX code path in the application.

 

[Edgemeal]

Pretty impressed with how my old chip is keeping up.

A 3930k at the same 4.2 Ghz only scores 32 points more.

I was expecting abit more with the IPC improvements SB has and access to the AVX code path in the application.

Try the latest Nightly build, I gained almost 5% on ivy i5.

The 4.2GHz battle

[B]
  Name           FPS         CPU       CPU MHz  Ram MHz  Timings      OS / CPU Capabilities[/B]
1 [URL="http://forums.anandtech.com/showpost.php?p=35215306&postcount=149"]YBS1[/URL]           354.352325  i7-3930K  4200       1866   -            MMX2 SSE2Fast SSSE3 SSE4.2 AVX                       
2 [URL="http://forums.anandtech.com/showpost.php?p=35202792&postcount=55"]SKORPI0[/URL]        324.455719  i7-3930K  4200          -   -            MMX2 SSE2Fast SSSE3 SSE4.2 AVX                       
3 [URL="http://forums.anandtech.com/showpost.php?p=35214137&postcount=142"]Makaveli[/URL]       319.169128  i7-970    4200       1600   7-8-7-20-1T  MMX2 SSE2Fast SSSE3 SSE4.2                           
4 [URL="http://forums.anandtech.com/showpost.php?p=35205269&postcount=78"]MrPickins[/URL]      253.857376  i5-4670K  4200       1600   -            MMX2 SSE2Fast SSSE3 SSE4.2 AVX AVX2 FMA3 LZCNT BMI2  
5 [URL="http://forums.anandtech.com/showpost.php?p=35213666&postcount=137"]videogames101[/URL]  221.463760  i5-3570K  4200       1600   9-9-9-24-1T  MMX2 SSE2Fast SSSE3 SSE4.2 AVX                       
6 [URL="http://forums.anandtech.com/showpost.php?p=35200979&postcount=16"]Edgemeal[/URL]       219.427078  i5-3570K  4200       1600   9-9-9-28-2T  MMX2 SSE2Fast SSSE3 SSE4.2 AVX                       
7 [URL="http://forums.anandtech.com/showpost.php?p=35213976&postcount=139"]gbeirn[/URL]         216.190186  i7-960    4200       1600   9-9-9-24-2T  MMX2 SSE2Fast SSSE3 SSE4.2                           
8 [URL="http://forums.anandtech.com/showpost.php?p=35201346&postcount=23"]nyker96[/URL]        194.348465  i5-2500K  4200       1600   -            MMX2 SSE2Fast SSSE3 SSE4.2 AVX

HandBreak Score Chart updated here (for now).

 

[moonbogg]

Great thread going here, and excellent job with the charts. I have to say though, that the 3930k seems pretty useless compared to the 4770k. Two less cores and damn near the same work produced. This is what sucks about the Intel release schedule. You get to pay top dollar for high end gear only to watch it be nullified by a newer gen mid range part shortly after. I guess thats the way it goes though.
Seriously though, is there anything that a 3930k is actually much better for than a 4770k? Anything at all?

 

[BallaTheFeared]

AVX2 is really helping Haswell in this, via the nightly.

I didn't even see Makaveli on the list, I never thought to look below my score...

Zing!

 

[blckgrffn]

Great thread going here, and excellent job with the charts. I have to say though, that the 3930k seems pretty useless compared to the 4770k. Two less cores and damn near the same work produced. This is what sucks about the Intel release schedule. You get to pay top dollar for high end gear only to watch it be nullified by a newer gen mid range part shortly after. I guess thats the way it goes though.
Seriously though, is there anything that a 3930k is actually much better for than a 4770k? Anything at all?

You got to enjoy Haswell performance in 2011? (fall release, right?)

I just bought a 3930k setup, I guess that makes me the fool

But man, if Intel wasn't releasing mid range parts that were eventually as fast as their high end stuff... well, that would be bad thing. Don't worry, given that you have to OC the snot out of Haswell CPU to come close to a 3930k with a medium OC, and we aren't going to see a significantly more powerful mainstream CPU until 2015 - at least - means the LGA2011 folks should be set for solid performance for years to come.

Plus more RAM capacity, more PCIe lanes, and making you more attractive - all benefits of LGA2011.

That is what I am banking on.

@ Balla - I am thinking off picking up another TF3 7950 for my rig. How is that working out for you?

 

[videogames101]

3570k @ 4.2Ghz
Memory: 1600 @ 9-9-9-24-1T

Nightly Build: HandBrake-5627svn-x86_64-Win_CLI

Using:

HandBrakeCLI --preset="Android" -i big_buck_bunny_480p_h264.mov -o big_buck_bunny_480p_h264.mp4 -v >log 2>&1

Output:
http://pastebin.com/YRVJWr8n

x264 [info]: using cpu capabilities: MMX2 SSE2Fast SSSE3 SSE4.2 AVX

[16:23:14] work: average encoding speed for job is 221.463760 fps

Note:
I've seen as high as [16:19:57] work: average encoding speed for job is 222.149155 fps, but I wasn't logging every run, so +/- 1 or 2 fps seems about right.

 

[cmdrdredd]

using cpu capabilities: MMX2 SSE2Fast SSSE3 SSE4.2 AVX
work: average encoding speed for job is 236.807419 fps

 

[gbeirn]

Going to post a few more:

HandBrake svn5622 (Nightly Build) - 64bit Version
OS: Microsoft Windows NT 6.1.7601 Service Pack 1
CPU: Intel(R) Core(TM) i7 CPU 960 @ 3.20GHz
Ram: 12278 MB, Screen: 1920x1080

RAM 1600 @ 9-9-9-24 2T For all runs

4.2GHz (21x200)
x264 [info]: using cpu capabilities: MMX2 SSE2Fast SSSE3 SSE4.2
x264 [info]: profile Main, level 3.0
[16:01:04] reader: done. 1 scr changes
[16:01:05] work: average encoding speed for job is 216.190186 fps

4.0GHz (20x200)
x264 [info]: using cpu capabilities: MMX2 SSE2Fast SSSE3 SSE4.2
x264 [info]: profile Main, level 3.0
[16:18:12] reader: done. 1 scr changes
[16:18:12] work: average encoding speed for job is 211.686844 fps

3.6GHz (18x200)
x264 [info]: using cpu capabilities: MMX2 SSE2Fast SSSE3 SSE4.2
x264 [info]: profile Main, level 3.0
[16:22:30] reader: done. 1 scr changes
[16:22:31] work: average encoding speed for job is 194.520248 fps

3.2GHz (24x133) Stock CPU
x264 [info]: using cpu capabilities: MMX2 SSE2Fast SSSE3 SSE4.2
x264 [info]: profile Main, level 3.0
[16:30:11] reader: done. 1 scr changes
[16:30:11] work: average encoding speed for job is 175.858643 fps

 

[BallaTheFeared]

@ Balla - I am thinking off picking up another TF3 7950 for my rig. How is that working out for you?

Hit or miss, sometimes awesome, sometimes fail, sometimes I get so tired of dinking with it I pull them and bench my HD4600.

 

[moonbogg]

You got to enjoy Haswell performance in 2011? (fall release, right?)

I just bought a 3930k setup, I guess that makes me the fool

But man, if Intel wasn't releasing mid range parts that were eventually as fast as their high end stuff... well, that would be bad thing. Don't worry, given that you have to OC the snot out of Haswell CPU to come close to a 3930k with a medium OC, and we aren't going to see a significantly more powerful mainstream CPU until 2015 - at least - means the LGA2011 folks should be set for solid performance for years to come.

Plus more RAM capacity, more PCIe lanes, and making you more attractive - all benefits of LGA2011.

That is what I am banking on.

@ Balla - I am thinking off picking up another TF3 7950 for my rig. How is that working out for you?

You know, you'r right. Having LGA2011 now is kind of like being Madonna in her 40's. We're not as hot as we used to be, but in a strange way still more desirable than our younger competitors.

 

[Makaveli]

AVX2 is really helping Haswell in this, via the nightly.

I didn't even see Makaveli on the list, I never thought to look below my score...

Zing!

Haswell does get a nice boost and I expect so with these encoding applications that are cpu dependant.

Ive been looking at everyone numbers some interesting scaling and boost with instruction sets

Not bad making the top 10 considering some of the crazy setups we have on this forum.

Try the latest Nightly build, I gained almost 5% on ivy i5.

The 4.2GHz battle

[B]
  Name           FPS         CPU       CPU MHz  Ram MHz  Timings      OS / CPU Capabilities[/B]
1 [URL="http://forums.anandtech.com/showpost.php?p=35202792&postcount=55"]SKORPI0[/URL]        324.455719  i7-3930K  4200          -   -            MMX2 SSE2Fast SSSE3 SSE4.2 AVX                       
2 [URL="http://forums.anandtech.com/showpost.php?p=35205697&postcount=84"]Makaveli[/URL]       292.537811  i7-970    4200       1600   7-8-7-20-1T  MMX2 SSE2Fast SSSE3 SSE4.2                           
3 [URL="http://forums.anandtech.com/showpost.php?p=35205269&postcount=78"]MrPickins[/URL]      253.857376  i5-4670K  4200       1600   -            MMX2 SSE2Fast SSSE3 SSE4.2 AVX AVX2 FMA3 LZCNT BMI2  
4 [URL="http://forums.anandtech.com/showpost.php?p=35213666&postcount=137"]videogames101[/URL]  221.463760  i5-3570K  4200       1600   9-9-9-24-1T  MMX2 SSE2Fast SSSE3 SSE4.2 AVX                       
5 [URL="http://forums.anandtech.com/showpost.php?p=35200979&postcount=16"]Edgemeal[/URL]       219.427078  i5-3570K  4200       1600   9-9-9-28-2T  MMX2 SSE2Fast SSSE3 SSE4.2 AVX                       
6 [URL="http://forums.anandtech.com/showpost.php?p=35213976&postcount=139"]gbeirn[/URL]         216.190186  i7-960    4200       1600   9-9-9-24-2T  MMX2 SSE2Fast SSSE3 SSE4.2                           
7 [URL="http://forums.anandtech.com/showpost.php?p=35201346&postcount=23"]nyker96[/URL]        194.348465  i5-2500K  4200       1600   -            MMX2 SSE2Fast SSSE3 SSE4.2 AVX

Wow so I just tested with the nightly build that is quite the boost.

HandBrake svn5622 (Nightly Build) - 64bit Version
OS: Microsoft Windows NT 6.1.7601 Service Pack 1
CPU: Intel(R) Core(TM) i7 CPU 970 @ 4.20GHz
Ram: 12279 MB, Screen: 1920x1200

x264 [info]: using cpu capabilities: MMX2 SSE2Fast SSSE3 SSE4.2
x264 [info]: profile Main, level 3.0
[18:33:07] reader: done. 1 scr changes
[18:33:08] work: average encoding speed for job is 319.169128 fps

 

[Makaveli]

Double post please delete

 

[Makaveli]

You know, you'r right. Having LGA2011 now is kind of like being Madonna in her 40's. We're not as hot as we used to be, but in a strange way still more desirable than our younger competitors.

And that desirable thing is Experience

young hot and dumb goes get boring after a while.

 

[Saffron]

I used my MSI GT70 Notebook:

i7-4700QM @ 2.4GHz (3.3GHz Boost)
16GB DDR3-1600
GTX770M

I get the following:

[15:32:10] starting job
[15:33:57] work: average encoding speed for job is 134.413071 fps

 

[BallaTheFeared]

319.169128 fps

:thumbsup:

 

[Jodell88]

I used my MSI GT70 Notebook:

i7-4700QM @ 2.4GHz (3.3GHz Boost)
16GB DDR3-1600
GTX770M

I get the following:

[15:32:10] starting job
[15:33:57] work: average encoding speed for job is 134.413071 fps

Something isn't right here. My sandy bridge mobile quad i7 and your haswell is even on performance. I have a lower clock speed than yours and don't have AVX2. What gives?

 

[Deders]

i5-750 @3800, Handbrake 0.9.9.5530, around 150FPS average on default android settings after a fresh install

http://pastebin.com/R8zyK9ZZ

 

[YBS1]

My results from a 4.2GHz run.

[01:09:13] work: average encoding speed for job is 354.352325 fps

HandBrake svn5622 (Nightly Build) - 64bit Version
OS: Microsoft Windows NT 6.2.9200.0
CPU: Intel(R) Core(TM) i7-3930K CPU @ 3.20GHz
Ram: 32709 MB, Screen: 1920x1080

 

[YBS1]

@4.9GHz

[01:45:29] work: average encoding speed for job is 413.095245 fps

HandBrake svn5622 (Nightly Build) - 64bit Version
OS: Microsoft Windows NT 6.2.9200.0
CPU: Intel(R) Core(TM) i7-3930K CPU @ 3.20GHz
Ram: 32709 MB, Screen: 1920x1080

Memory is 1866 btw...