What Handbrake encode performance do you get (mainly i5 2500K users but others welcome)

titaniumx3 · 15 Dec 2011 at 14:14

I meant in general it's wise to get people to verify their results. Not saying anyone would, but you never know!

moogle · 15 Dec 2011 at 19:10

Not an i5 2500K but I wanna play!

Q9550 @ 4GHz

Don't know if I did it right but I imported the plist file and then pressed start. Was that right?

Actually I might have done it wrong, the settings are at 15mbps and it's not resizing to 720p. Do I need a newer version of Handbrake to import that?

arknor · 15 Dec 2011 at 19:25

1055t @ stock

buahahaha beast some of the i7 ****K users

pgi947 · 15 Dec 2011 at 20:11

2600k @ 4.8ghz

Agree that a larger file (30sec-1min) would give better results.

Edit: did the same as orcvader, moved source and output to a another drive and...

drunkenmaster · 15 Dec 2011 at 22:44

Guys, you can enable logging and get the final average

PCZ · 16 Dec 2011 at 08:45

John24 said:
Would anyone be willing to do a few encodes with Handbrake?

If so you can download handbrake here.
Use the Normal Preset.
Set the Contrainer to MKV.
Set the Priority Level to High in the options.

When finnished each encode record the average FPS, select the Activity Window then Encode Log from the dropdown, the average FPS will be near the end.

Source files, the smallest one would be enough but if you can do all 3 that would be better.

Tron_Trailer_HD.zip - 268 MB
Splice_Trlr_1_1080P.wmv - 336 MB - Right Click Save As this one.
big_buck_bunny_1080p_surround.avi - 886 MB

My 1100 results for comparison.

Code:

1100 24/7 OC 4.0GHz | 3.0 | 1666-9-9-9-24-34 Tron 43.162312 fps Splice 29.393965 fps Big 43.118412 fps

EDIT: Handbrake will max all your cores so keep an eye on temps, I don't want to be responsible for frying your lovely new chip

Cheers.

The file you are testing with is far too small.
most of the short time that encode is taking is reading and writing to the HD.
See above instructions for a better encoding test.

mrk · 16 Dec 2011 at 09:02

Those are not BluRay sources either though, you'd be transcoding from one lossy format to a near lossless format, my Op was referring to transcoding actual BluRay sources to mkv contained h264 videos!

PCZ · 16 Dec 2011 at 09:51

mrk said:
Those are not BluRay sources either though, you'd be transcoding from one lossy format to a near lossless format, my Op was referring to transcoding actual BluRay sources to mkv contained h264 videos!

Thats fine but blue ray sources are very large.
Not suitable for a comparison benchmark that members can easily do and post results.

Folks would need the BD disc to rip and then encode.
The time taken to encode would vary a bit between different sources, so you don't want people posting encoding times from different rips.

I have a variety of PC's here and can run comparison benchmarks using the same rip if you want.
Obviously it would be some royalty free content

FredFlint · 16 Dec 2011 at 15:29

Just ran it on my 2760m, was 29 fps.

titaniumx3 · 17 Dec 2011 at 00:46

mrk said:
Those are not BluRay sources either though, you'd be transcoding from one lossy format to a near lossless format, my Op was referring to transcoding actual BluRay sources to mkv contained h264 videos!

Relative performance between CPUs would be more or less the same though as the native BluRay format is using lossless encoding anyway, just less aggressive and at higher bitrate.

mrk · 17 Dec 2011 at 00:50

Hmm so a requirement for a standard profile and standard media file large enough to get a decent average figure at the end of it.

Hmmmm.

1337z0r · 17 Dec 2011 at 09:08

I'm doing some encodoing now with a larger bluray source if my own going to 720 and , I'm getting 60fps and this is going to a 5400rpm drive. I fairly confiedent write speeds can also play a big part in taking these cpu's to full potential.

My 2700 is running at about 50 - 70 percent, the only bottleneck can be write speed.

DragonQ · 17 Dec 2011 at 12:31

In theory, an i7-920/930 @ 4 GHz should be the same, if not slightly better, than an i5-2500K @ 4.5 GHz. Unfortunately nearly all of the results in this thread are from i7-2600Ks...

Anyway, I get 39.175694 fps on the Tron trailer.

PCZ · 18 Dec 2011 at 04:03

Ran the tests on a variety of PC's i have around the house.
I did post some of these reults in another thread but this list contains a few more machines for comparison.

Handbrake 0.9.5
Normal Preset
Container MKV
Priority high

3930K 4500Mhz
RAM 1600 9-9-9-24
Tron 65.488953 fps
Splice 46.229198 fps
Big 63.719532 fps

2600K 4600Mhz
RAM 1600 9-9-9-24
Tron 51.616753 fps
Splice 35.502117 fps
Big 51.450516 fps

FX8150 4600Mhz
RAM 1600 9-9-9-24
Tron 48.685398 fps
Splice 34.118172 fps
Big 47.157837 fps

2500K 4600Mhz
RAM 1600 9-9-9-24
Tron 44.214649 fps
Splice 30.144783 fps
Big 45.051327 fps

I7 920 4000Mhz
RAM 1600 9-9-9-24
Tron 42.470345 fps
Splice 29.149643 fps
Big 42.311901 fps

X6 1090T 4000Mhz
RAM 1600 9-9-9-24 NB 2800Mhz
Tron 42.315029 fps
Splice 28.868134 fps
Big 42.403152 fps

I5 750 4000Mhz
RAM 1600 9-9-9-24
Tron 34.977623 fps
Splice 23.885386 fps
Big 34.408932 fps

X4 965 4000Mhz
RAM 1600 9-9-9-24 NB 2400Mhz
Tron 29.127506 fps
Splice 20.053186 fps
Big 29.122873 fps

The 3930k SB-E is sitting on top, but it jolly well should be considering how much it cost. (XMAS present to myself)

The 2500K appears quite a bit slower than the 2600K at encoding.
HyperThreading is showing it's worth here giving the 2600K quite an advantage over the 2500K.

The FX8150 slotted in between the 2500k and 2600K.
I expected better of a supposed 8 core CPU, best we leave it at that.

The I7 920 and the X6 1090T did well considering there age and comparatively low clock speeds, being almost as quick as the 2500k running 600Mhz faster.

pgi947 · 18 Dec 2011 at 10:02

Wow look at those 12 threads of sb-e go. The 8core bdfx gets a little more pathetic every time I see it.

titaniumx3 · 18 Dec 2011 at 17:10

Using the video files and encoding settings posted by PCZ, I get the following average speeds on a 2700K @ 4.6ghz, 8GB DDR3 1600mhz (9-9-9-24 1T):

Tron: 51.92 fps
Splice: 35.52 fps
Big: 51.94 fps

Seems consistent with PCZ's results. I did notice that it's not using the CPU fully as the CPU utilisation frequently drops down to the 60-80% range, though it's mostly at 99%. I did put the source files on my SSD and the encoded file destination on a mechanical drive, though I doubt very much that it makes any difference; you'd have to be encoding at high speed with high bitrate settings to saturate the read-write throughput of modern mechanical hards disks, let alone SSDs.

From the results it seems at least in handbrake, going from Core i5 SB to Core i7 SB gives a 15-20% improvement (clock for clock) - which is pretty much what you'd expect.

Bossk128 · 11 Jan 2016 at 15:54

Resurrection saves a new thread

I currently use an I5 2400 and get 18fps transcosing Blu Ray rips to MKV using High, 4.1, pass through audio.

My motherboard will support an I7 (2600(k), but I want to know roughly what speed increase I'll get moving from 3.1Ghz quad core to 3.4Ghz HyperThreaded 4 core.

Anybody done the test?

cheers.

wellibob · 12 Jan 2016 at 14:10

In my experience the hyperthreading works wonders with handbrake. Speeds will vary but a i7 or above over a 4 core i5 anyday for encoding.

Bossk128 · 12 Jan 2016 at 15:10

Nice one thanks.

Slightly OT, does anyone know if an I5 2400 heat sink is compatible (by this I mean draws off heat fast enough) with an I7 2600/2700 at stock?

Orcvader · 12 Jan 2016 at 17:51

I think they're even the same heatsink. It does the job, but don't you find it loud

? What motherboard do you have? The 2600K can easily overclock to 4GHz on stock volts. The i7-3770K may also be compatible with your board (well, it is with my Z68 board).