Benchmark your computer @4K with Handbrake 1.1 and H265! NOW WITH ZEN3!

[ElFenix]

edit: now with HB 1.3.3 results and zen 3 is a monster:

Page 14 - Benchmark your computer @4K with Handbrake 1.1 and H265! NOW WITH ZEN3!

Page 14 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

forums.anandtech.com

new testing regime:

If you're up for it let's try it with a more rigorous approach. Instructions below if you want to post a score.

Rank	User	CPU	Handbrake Version Verified	Total CPU Usage (Hwinfo)	Average Effective Clockspeed (Hwinfo)	seconds	fps/core/GHz (fps per available GHz)	fps	CPU Core	Cores
1	EvilRick	Ryzen 7 2700x 8/16	1.3.3	95.2%	3.910	392	0.147	4.61	Zen+	8
2	Hulk	4770K 4/8	1.3.3	98.1%	3.620	815	0.153	2.22	Haswell	4
3	Hulk	8250u 4/8 (Surface 2 laptop)	1.3.3	98.3%	2.400	1065	0.177	1.70	Kaby Lake R	4

1. Use the following version of Handbrake with the built-in h.265 mkv 2160p60 preset
HandBrake-1.3.3-x86_64-Win_GUI.exe
Once this current version is replaced you'll be able to access it from the following link.
HandBrake: Nightly Builds
Nightly builds of HandBrake

HandBrake: Old Releases

Old Releases of HandBrake

handbrake.fr

2. Report average effective clockspeed using HWinfo - https://www.hwinfo.com/
I'm sure there are better ways to do this using the logging feature but here's how I did it. I downloaded the portable version, which runs from the exe file.

Run the exe and select "sensors only." You only need this window open.
Minimize the sensors window so you can have it and Handbrake visible at the same time. Select the "average effective clockspeed" line so you can spot it easily. It will highlight.
Note the icon for "reset min/max/average values and timer" in the bottom right area of the window, it's the clock icon.

Start the handbrake but don't start/reset the timer until handbrake is actually encoding video. Record the average effective clockspeed right before handbrake finishes the encode. It should settle in pretty quickly after the encode starts and not change during the encode.

3. Report "Total CPU Usage" (also from HWinfo)

4. Report your encoding time and CPU model

edit: results chart from @Hulk :

Page 12 - Benchmark your computer @4K with Handbrake 1.1 and H265! NOW WITH ZEN3!

Page 12 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

forums.anandtech.com

edit:
4k HDR sample now on google drive:

et voila. we can all host the file and round robin it

LG New York HDR UHD 4K Demo.zip

drive.google.com

Because many-core processors have low utilization doing 1080p recode, lets stress them out using 4k. Our new sample is a 4K HDR sample used to sell TVs.

For this test we'll be using handbrake 1.1.0 (1.1.1 has been released, please use 1.1.0 from the linked archive) and its built-in h.265 mkv 2160p60 preset:

(if you're upgrading and don't see that preset, select "Reset Built-In Presets" under the Preset menu)

As before, go into Handbrake's Preferences, make sure QuickSync Decoding is unchecked under Video, and check the box to put a copy of individual encode logs either with the same location or in a specified location.

The line we're looking for is near the bottom of the log and looks like this:

encoded 1806 frames in 1709.87s (1.06 fps), 11820.04 kb/s, Avg QP:29.09

Report your results with your processor, stock/overclock speed, operating system, and ram speed/amount. I'll pull the results into a spreadsheet and update this OP with the leaderboard from time to time.

 

[Markfw]

encoded 1806 frames in 394.29s (4.58 fps), 11820.04 kb/s, Avg QP:29.09

E5-2696v3 ES (I am pretty sure) 14 cores/28 threads@2.5 ghz

Windows 10, ram is 16 gig of 2133

 

[Markfw]

encoded 1806 frames in 242.49s (7.45 fps), 11820.04 kb/s, Avg QP:29.09

1950x 16c/32t @3.8

Windows 10 32 gig PC3200 cl14

 

[Markfw]

BTW, on both computers, after running the test and exiting the application, boinc would not go over 50% CPU usage. I had to reboot the computer to get things back to normal.

VERY odd

 

[Hail The Brain Slug]

5960x at 4.7GHz core, 4.6GHz cache, DDR4 3200 14-15-15-30-1T

Windows 10 Pro

encoded 1806 frames in 319.84s (5.65 fps), 11820.01 kb/s, Avg QP:29.09

Not bad. I'll be killing some applications and doing additional runs to see if I can shave off a few seconds.

 

[IEC]

8700K at 4.7GHz all core, 4.7GHz cache, no AVX offset, -50mV vCore offset, DDR4-3600 CL15 @ XMP timings + pretty much all power limits maxed out and power saving disabled + Windows 10 Pro latest build w/High Performance power option.

encoded 1806 frames in 359.41s (5.02 fps), 11820.04 kb/s, Avg QP:29.09

 

[epsilon84]

encoded 1806 frames in 340.42s (5.31 fps), 11820.04 kb/s, Avg QP:29.09

8700K @ 5.0GHz / 4.9GHz AVX offset, 16GB DDR4 3400 15-16-16-35-2T running Windows 10 Pro

 

[SlimFan]

encoded 1806 frames in 2631.13s (0.69 fps), 11820.04 kb/s, Avg QP:29.09

Gemini Lake J5005 NUC.

 

[cmdrdredd]

encoded 1806 frames in 386.65s (4.67 fps), 11820.04 kb/s, Avg QP:29.09

i7 8700k 4.7Ghz w/4.5Ghz AVX offset. 16GB Corsair LPX DDR4-3200 @ 16-19-19-36

I don't have the best overclocking chip and I did not want to bother delidding it. My AVX offset has to be 4.5Ghz for stability and my score reflects that. This is mostly a gaming system and I don't use AVX enabled programs regularly so it's fine for me. A couple hundred Mhz isn't a deal breaker at all.

 

[mjdupuis]

encoded 1806 frames in 520.30s (3.47 fps), 11820.04 kb/s, Avg QP:29.09
Win 10 Pro
Ryzen 2600X at stock clocks, 16GB GSkill Trident Z @ 3466 14-14-14-30

 

[jones377]

encoded 1806 frames in 768.63s (2.35 fps), 11820.04 kb/s, Avg QP:29.09

i7 4770K 4C/8T Haswell at stock. 16GB DDR3 1600 9-9-9-24. Win10 Pro

 

[gamervivek]

encoded 1806 frames in 593.10s (3.05 fps), 11820.04 kb/s, Avg QP:29.09

ryzen 1600@3.5Ghz. 64GB DDR4 2666 14-14-14-35, Win 10 Home

 

[Hail The Brain Slug]

encoded 1806 frames in 768.63s (2.35 fps), 11820.04 kb/s, Avg QP:29.09

i7 4770K 4C/8T Haswell at stock. 16GB DDR3 1600 9-9-9-24. Win10 Pro

By "at stock" do you mean full stock with tdp limits throttling your core speed down to ~3.5 GHz or do you mean multicore enhanced 3.9 Ghz under load?

 

[Hitman928]

encoded 1806 frames in 242.49s (7.45 fps), 11820.04 kb/s, Avg QP:29.09

1950x 16c/32t @3.8

Windows 10 32 gig PC3200 cl14

Can you post a snapshot of thread usage while running the test? I'm curious to see how many cores are actually being used.

 

[cfenton]

I ran the benchmark on a few machines. I thought it would be interesting to see how some of the older and cheaper ones manage.

1. encoded 1806 frames in 476.90s (3.79 fps), 11820.04 kb/s, Avg QP:29.09 - Ryzen 1700 @ 3.7Ghz, Windows 10, 16GB DDR4 2933
2. encoded 1806 frames in 5099.43s (0.35 fps), 11820.04 kb/s, Avg QP:29.09 - Chuwi Lapbook 14.1, Celeron N3450 stock (1.1Ghz turbo to 2.2Ghz), Linux Mint 18.3 KDE, 4GB DDR3 1600
3. encoded 1806 frames in 6027.14s (0.30 fps), 11820.04 kb/s, Avg QP:29.09 - Celeron 2957U stock (1.4Ghz no turbo), Linux Mint 18.3 KDE, 6GB DDR3 1333
4. encoded 1806 frames in 1586.81s (1.14 fps), 11820.04 kb/s, Avg QP:29.09 - I5 3570s stock (3.1 Ghz turbo to 3.6Ghz), Windows 10, 10GB DDR3 1333

 

[ChippyUK]

Intel i5-8400 - Windows 10, 16GB DDR4 3200C16 (All cores at 100% during encoding process)

encoded 1806 frames in 558.91s (3.23 fps), 11820.04 kb/s, Avg QP:29.09

 

[lightmanek]

encoded 1806 frames in 414.34s (4.36 fps), 11820.04 kb/s, Avg QP:29.09

Ryzen 7 2700X @4ghz all cores, 16GB DDR 3200 CL14-13-13 LL, Windows 10 64bit

 

[Bubbleawsome]

i5 7300HQ 2.5/3.5 with an all core turbo of 3.2Ghz. 8GB RAM probably 2400Mhz, Windows 10 64-bit

encoded 1806 frames in 1022.42s (1.77 fps), 11820.04 kb/s, Avg QP:29.09

 

[.vodka]

encoded 1806 frames in 430.69s (4.19 fps), 11820.04 kb/s, Avg QP:29.09

R7 1700 - 3800MHz @ 1.32v LLC0, relaxed EDC on, performance bias on, DDR4-3466 15-15-15-35 1T + Stilt's 3466MHz profile. Daily settings

----------------------------------------------------------

encoded 1806 frames in 416.04s (4.34 fps), 11820.04 kb/s, Avg QP:29.09

R7 1700 - 3950MHz @ 1.43v LLC4, relaxed EDC on, performance bias on, DDR4-3466 15-15-15-35 1T + Stilt's 3466MHz profile. As far as it'll go

---------------------------------------------------------------------------

Man, that voltage wall is more present than ever on this very early R7 1700, batch 1708
3975-4025MHz would crash at 1.45v + LLC4 set in BIOS, had to back down to 3950MHz with all the performance enhancing options on to finish the encode.. and that was at ~80°C on this and ~18°C ambient temperature. Need custom WC to go on.

Without these tweaks (relaxed EDC and performance bias) sure, you can clock higher at the same voltage but then you end up at about the same final performance in this test so I leave them on. On other tests...

CPU usage is now 90-100%, a 1080p H265 encode as used in the thread before this one would be at ~40-50% CPU usage.

 

[eek2121]

Threadripper 1950X @ 4.1 GHz, RAM DDR4 3200 CL16 (32 GB)
Windows 10 with all the latest patches

encoded 1806 frames in 254.04s (7.11 fps), 11820.04 kb/s, Avg QP:29.09

 

[Despoiler]

encoded 1806 frames in 444.03s (4.07 fps), 11820.04 kb/s, Avg QP:29.09

Ryzen 1700 @ 3.9ghz 3200 CL 14 1.387v auto LLC

 

[ElFenix]

 

[richaron]

Ubuntu 18.04 (4.15.0-20-generic)
Ryzen 1700 @ 3.7 (1.2V in BIOS, 1.25-1.27V reality? [monitoring is hard])
64GB ECC @ 3000 CL17 (4 x 16GB dual ranked @ 1.35V)

encoded 1806 frames in 434.68s (4.15 fps), 11725.82 kb/s, Avg QP:29.05

 

[mjdupuis]

encoded 1806 frames in 347.50s (5.20 fps), 11820.04 kb/s, Avg QP:29.09

Win 10 Pro
8700K @ 4.8 (no AVX offset), 16GB Trident Z @ 14-14-14-28-1T

 

[JoeRambo]

encoded 1806 frames in 342.38s (5.27 fps), 11820.04 kb/s, Avg QP:29.09

8700K @ 4.8, 4.4 uncore, mem @3200 13-15-15-32-1T