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

Fir · Mar 7, 2020

I can but more concerning is the NUMA mode that can't go away! On the most up to date Win10 Enterprise it shows single socket and L3 cache properly. I put 2019 Server Datacenter on another drive and ran updates until it says it's up to date and that actually shows it as a dual socket system!

Here is the selection from the task manager

There is a 0 option in the BIOS which should disable NUMA but it doesn't change this behavior. Under 'set affinity' options there are two processing groups of 64 cores and by default it only shows it using one group. The other group is assignable but affinity isn't always settable by the user - even admin.

I did run some blender files on the 10 Enterprise install and those render times did scale quite nicely. Socket power use over 700W according to HWinfo with average core close to 4.1GHz. Temp in the 50s. Using a Heatkiller IV solid copper block that's piped into a chilled water loop maintained at 18°C. That tap is off a system with about 30 tons capacity so there's room for some more o/c. I'd like to see what 4.4GHz across all cores is capable of!

DrMrLordX · Mar 7, 2020

@Fir

Are you having the same problem under Linux?

Yeroon · Mar 7, 2020

Fir said:
I can but more concerning is the NUMA mode that can't go away! On the most up to date Win10 Enterprise it shows single socket and L3 cache properly. I put 2019 Server Datacenter on another drive and ran updates until it says it's up to date and that actually shows it as a dual socket system!

I do believe this a windows problem and not a bios setting based on the 3990x review here on anandtech.
Handbrake is available on linux if you want to see how well that 3990 scales.

Hulk · Mar 7, 2020

Rank	User	CPU	GHz	seconds	fps/core/GHz	fps	Architecture	Cores
1	Fir	3990x 64/128	4.10	94.86	0.073	19.04	Zen 2	64
2	Fir	3970x 32/64	4.15	105.63	0.129	17.10	Zen 2	32
3	DaPoets	3960X 24/48	4.00	123.66	0.152	14.60	Zen 2	24
4	Charlie22911*	7980XE 18/36	4.37	141.12	0.163	12.80	Skylake	18
5	YBS1*	3950x 16/32	3.50	154.72	0.208	11.67	Zen 2	16
6	Micromd	2 - Xeon 8260L ES 48/96	2.80	156.99	0.086	11.50	Cascade Lake	48
7	Markfw	EPYC 7742 64/128	3.00	162.97	0.058	11.08	Zen	64
8	Bavor	3950x 16/32	4.05	169.73	0.164	10.64	Zen 2	16
9	Kjboughton	E5-2696 2x18 36/72	3.99	174.06	0.072	10.38	Broadwell	36
10	DrMrLordX	3900x 12/24	4.35	191.78	0.180	9.42	Zen 2	12
11	rvborgh	Quad Opteron (48xK10)	3.00	200.00	0.063	9.03	Piledriver	48
12	Muadib	3900x 12/24	4.30	202.15	0.173	8.93	Zen 2	12
13	Cata40	7840xe 18/36	3.40	211.00	0.140	8.56	Skylake	18
14	Fir	2990WX 32/64	4.10	216.35	0.064	8.35	Zen+	32
15	Markfw	1950x 16/32	3.80	242.49	0.122	7.45	Zen	16
16	mrpiggy	E5-2699 18/36	3.70	253.41	0.107	7.13	Haswell	18
17	eek2121	1950x 16/32	4.10	254.04	0.108	7.11	Zen	16
18	RobertR1	9900k 8/16 (52C/47UC)	5.20	254.98	0.170	7.08	Coffee Lake	8
19	Lexingtonian	Quad Opteron 6386	3.50	275.66	0.029	6.55	Piledriver	64
20	Kenmitch*	Ryzen 7 3700x 8/16	3.97	279.43	0.203	6.46	Zen 2	8
21	CricketSP	9900k 8/16	4.90	281.80	0.163	6.41	Coffee Lake	8
22	kmm0000	9900k 8/16	4.60	290.96	0.169	6.21	Coffee Lake	8
23	Makaveli	3800x 8/16	4.90	292.40	0.158	6.18	Zen 2	8
24	TillausHamburg	Xeon ES QL1L 20/40	2.00	297.54	0.152	6.07	Skylake SP	20
25	sdifox	3700x 8/16	4.00	306.52	0.184	5.89	Zen 2	8
26	Chaotic42	9700k 8/8	5.00	311.03	0.145	5.81	Coffee Lake	8
27	XabanakFanatik	5960x 8/16	4.70	319.84	0.150	5.65	Haswell	8
28	DooKey	7820X 8/16	4.50	320.75	0.156	5.63	Skylake	8
29	epsilon 84	8700K 8/16 (5/4.9)	5.00	340.42	0.133	5.31	Coffee Lake	8
30	JoeRambo	8700k 8/16 (4.8/4.4)	4.80	342.38	0.137	5.27	Coffee Lake	8
31	mjdupuis	8700k 8/16 (4.8/4.8)	4.80	347.50	0.135	5.20	Coffee Lake	8
32	IEC	8700K 8/16 (4.7/4.7)	4.70	359.41	0.134	5.02	Coffee Lake	8
33	aclos3	3600x 6/12	4.04	367.20	0.203	4.92	Zen 2	6
34	CakeMonster	8700k 8/16 (4.9/4.6)	4.90	379.09	0.122	4.76	Coffee Lake	8
35	cmdrdredd	8700K 8/16 (4.7/4.5)	4.70	386.65	0.124	4.67	Coffee Lake	8
36	rgba	8700 8/16	4.30	387.57	0.135	4.66	Coffee Lake	8
37	Markfw	E5-2696v3 14/28	2.50	394.29	0.131	4.58	Broadwell	14
38	Yeroon	Ryzen 3600 6/12	3.60	402.25	0.208	4.49	Zen 2	6
39	Markfw	2700x 8/16	4.10	404.44	0.136	4.47	Zen+	8
40	Dan057	Ryzen 2700 8/16	4.00	410.08	0.138	4.40	Zen+	8
41	lightmanek	Ryzen 7 2700x 8/16	4.00	414.34	0.136	4.36	Zen+	8
42	.vodka	Ryzen 1700x 8/16	3.95	416.04	0.137	4.34	Zen	8
43	pjmssn	Dual Xeon E5-2690 16/32	2.90	416.47	0.093	4.34	Sandy Bridge EP	16
44	Bburninatortech4	2700x 8/16	4.00	423.90	0.133	4.26	Zen+	8
45	Franz316	2700x 8/16	3.80	429.95	0.138	4.20	Zen+	8
46	AnnoyedGrunt	Ryzen 7 2700x 8/16	4.10	430.94	0.128	4.19	Zen+	8
47	richaron	Ryzen 1700 6/12 (Ubuntu 18.04)	3.70	434.68	0.187	4.15	Zen	6
48	CrackRabbit	2700x 8/16	3.80	440.29	0.135	4.10	Zen+	8
49	Despoiler	Ryzen 1700 8/16	3.90	444.03	0.130	4.07	Zen	8
50	ArisVir*	2700 8/16	3.80	454.14	0.131	3.98	Zen+	8
51	dfenton	Ryzen 1700 8/16	3.70	476.90	0.128	3.79	Zen	8
52	CuriousMike	9600k 6/6	4.40	477.30	0.143	3.78	Coffee Lake	6
53	Xenon	Ryzen 5 1600AF 6/12	4.00	504.29	0.149	3.58	Zen+	6
54	monkeydelmagico*	Ryzen 1600 AF 6/12	4.09	504.45	0.146	3.58	Zen	6
55	William Gaatjes	Ryzen 2600 6/12 (memory 3200)	3.60	519.91	0.161	3.47	Zen	6
56	mjdupuis	Ryzen 2600x 6/12	4.20	520.30	0.138	3.47	Zen+	6
57	ChippyUK	i5-8400 6/6	4.00	558.91	0.135	3.23	Coffee Lake	6
58	ywasmd	i5 8400 6/6	4.00	562.68	0.134	3.21	Coffee Lake	6
59	gamervivek	Ryzen 1600 6/12	3.50	593.10	0.145	3.05	Zen	6
60	CronoReverse	4790k 4/8	4.50	644.52	0.156	2.80	Haswell	4
61	Hulk*	4770K 4/8	3.90	754.96	0.153	2.39	Haswell	4
62	jones377	4770K 4/8	3.90	768.63	0.151	2.35	Haswell	4
63	Nathan Gall	4770K 4/8	3.90	775.24	0.149	2.33	Haswell	4
64	Junky228	Xeon w3690 6/12	4.00	877.63	0.086	2.06	Westmere EP	6
65	Imported_stryfe	2600k 4/8	4.50	951.72	0.105	1.90	Sandy Bridge	4
66	Bubbleawsome	i5-7300HQ (laptop) 4/4	3.20	1022.42	0.138	1.77	Kaby Lake	4
67	Hulk*	8250u 4/8 (Surface 2 laptop)	2.50	1067.13	0.169	1.69	Kaby Lake R	4
68	NickatNight1980	FX8320 4/8	4.16	1117.63	0.097	1.62	Piledriver	4
69	Xenon	i5-1035G4 4/8	2.10	1126.01	0.191	1.60	Ice Lake	4
70	Xenon	i5-3570k 4/4	4.00	1206.06	0.094	1.50	Ivy Bridge	4
71	Xenon*	FX-6300 6/6	4.10	1336.18	0.055	1.35	Piledriver	6
72	mylesrom	i5-3570 4/4	3.80	1399.56	0.085	1.29	Ivy Bridge	4
73	dfenton	i5-3570s 4/4	3.60	1586.81	0.079	1.14	Ivy Bridge	4
74	Imported_stryfe	i5-750 4/4	2.80	2012.80	0.080	0.90	Lynnfield	4
75	Slim Fan	Pentium Silver J5005 (NUK) 4/4	2.80	2631.13	0.061	0.69	Goldmont Plus	4
76	PeterScott	Core 2 Quad Q9400 4/4	3.20	2655.10	0.053	0.68	Yorkfield	4
77	ywasmd	i5 3210 (Macbook Pro) 2/4	3.10	3144.79	0.093	0.57	Ivy Bridge	2
77	Fanatical Meat	Q9650 4/4	3.00	3439.20	0.044	0.53	Yorkfield	4
78	dfenton	Celeron N3450 (laptop) 4/4	2.20	5099.43	0.040	0.35	Apollo Lake	4
79	dfenton	Celeron 2957U 2/2	1.40	6027.14	0.107	0.30	Haswell	2

Fir · Mar 8, 2020

Yeroon said:
I do believe this a windows problem and not a bios setting based on the 3990x review here on anandtech.
Handbrake is available on linux if you want to see how well that 3990 scales.

I'm familiar with Linux on controllers, servers, etc. (CLI only)
I'm guessing I'd need to use a GUI based distro.
I am open to any suggestions on distros to try. Got plenty of bandwidth and space to burn!

Yeroon · Mar 8, 2020

Fir said:
I'm familiar with Linux on controllers, servers, etc. (CLI only)
I'm guessing I'd need to use a GUI based distro.
I am open to any suggestions on distros to try. Got plenty of bandwidth and space to burn!

If you're comfortable with CLI, you might be able to get away without a gui:

HandBrake Documentation — CLI Options

Now, I am not great at it so cant say it will work 100%, but it does seem like its possible.

As for using a GUI linux, I'm on Linux Lite, which is based on ubuntu 18.04 LTS/Debian. Its got a windows pre win8 style start/feel, but I like it since its a fairly lean install. I'm sure theres a gui based version of whatever command line you are comfortable with.

Gideon · Mar 9, 2020

Fir said:
I'm familiar with Linux on controllers, servers, etc. (CLI only)
I'm guessing I'd need to use a GUI based distro.
I am open to any suggestions on distros to try. Got plenty of bandwidth and space to burn!

Oddly it seems that Intel's Clear Linux might be the fastest distro to try out (comparison to other distros):

Benchmarking 9 Linux Distributions On A $50 Processor - Phoronix

www.phoronix.com

And with 3990x in particular:

Benchmarking 9 Linux Distributions On A $50 Processor - Phoronix

www.phoronix.com

Fir · Mar 12, 2020

Ok up and running on Intel(r) Clear Linux Project.

Here's what I found:

encoded 1806 frames in 111.36s (16.22 fps), 11820.04 kb/s, Avg QP:29.09

[01:24:24] CPU: AMD Ryzen Threadripper 3990X 64-Core Processor
[01:24:24] - logical processor count: 64

Shouldn't logical processor count show 128?

As you can see, it's considerable worse than Win10!

DrMrLordX · Mar 12, 2020

Weird. It's only utilizing half the CPU.

Markfw · Mar 12, 2020

DrMrLordX said:
Weird. It's only utilizing half the CPU.

That happened to me too.

Yeroon · Mar 12, 2020

Fir said:
Ok up and running on Intel(r) Clear Linux Project.

Here's what I found:

encoded 1806 frames in 111.36s (16.22 fps), 11820.04 kb/s, Avg QP:29.09

[01:24:24] CPU: AMD Ryzen Threadripper 3990X 64-Core Processor
[01:24:24] - logical processor count: 64

Shouldn't logical processor count show 128?

As you can see, it's considerable worse than Win10!

Didn't expect it to only see 1/2 your threads on linux, though after some reading it seems the x265 encoder is limited to 64 threads unless told otherwise:
https://x265.readthedocs.io/en/default/cli.html (check out the pools info section)

Also, from the log file :x265 [info]: Coding QT: max CU size, min CU size : 64 / 8

You can possibly add options for more threads in the video tab of handbrake, in the more settings window.
pools=n:frame-threads=n

I was not able to get it to use less than 12 threads on my 3600, but since I dont have 64 cores I cant figure out if its possible to get it to use more.

Fir · Mar 20, 2020

Best on Windows 10 1909 so far

HandBrake 1.3.1 (2020010400)
OS: Microsoft Windows NT 10.0.18363.0
CPU: AMD Ryzen Threadripper 3990X 64-Core Processor
Ram: 261888 MB,
GPU Information:
NVIDIA Quadro RTX 8000 - 26.21.14.4250
Screen: 7680×4320

encoded 1806 frames in 94.28s (19.15 fps), 11820.04 kb/s, Avg QP:29.09

vernieman · Jun 7, 2020

Fir said:
Best on Windows 10 1909 so far

HandBrake 1.3.1 (2020010400)
OS: Microsoft Windows NT 10.0.18363.0
CPU: AMD Ryzen Threadripper 3990X 64-Core Processor
Ram: 261888 MB,
GPU Information:
NVIDIA Quadro RTX 8000 - 26.21.14.4250
Screen: 7680×4320

encoded 1806 frames in 94.28s (19.15 fps), 11820.04 kb/s, Avg QP:29.09

64 cores 🤤🤤🤤

vernieman · Jun 7, 2020

HandBrake 1.3.2 (2020050300)
OS: Microsoft Windows NT 10.0.18363.0
CPU: Intel(R) Core(TM) i5-10600K CPU @ 4.10GHz
Ram: 16305 MB,
GPU Information:
NVIDIA GeForce GTX 1060 6GB - 26.21.14.4614
Screen: 2560x1440

encoded 1806 frames in 398.30s (4.53 fps), 11820.04 kb/s, Avg QP:29.09

HairNutz · Oct 7, 2020

Windows 10 Pro
Build 19041.508
Handbrake x64 1.1.0

i7 8700k @ stock clocks 3.7ghz / 4.7ghz turbo
16gb quad channel ddr4 @ 3000MHz
encoded 1806 frames in 402.55s (4.49 fps), 11820.04 kb/s, Avg QP:29.09

i7 8700k OC @ 4.9ghz / undervolted 1.28
16gb quad channel ddr4 @ 3000MHz
encoded 1806 frames in 360.37s (5.01 fps), 11820.04 kb/s, Avg QP:29.09

.vodka · Oct 7, 2020

.vodka said:
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

Updating my entry.

3900x, PBO +100MHz, stock PPT TDC EDC values, -0.075mV undervolt, LLC auto
encoded 1806 frames in 200.82s (8.99 fps), 11820.04 kb/s, Avg QP:29.09

3900x, per CCX OC (4375, 4375, 4325, 4300) 1.25v LLC3 as found by 1usmus' CTR tool
encoded 1806 frames in 190.61s (9.47 fps), 11820.04 kb/s, Avg QP:29.09

Both cases, 1900MHz fclk, 4x16GB DDR4-3800 16-20-16-16-36-56 1T+GDM, tuned secondaries and tertiaries

Twice as fast over my old maxed out 1700. Not too bad!

moinmoin · Oct 7, 2020

.vodka said:
3900x, PBO +100MHz, stock PPT TDC EDC values, -0.075mV undervolt, LLC auto
encoded 1806 frames in 200.82s (8.99 fps), 11820.04 kb/s, Avg QP:29.09

3900x, per CCX OC (4375, 4375, 4325, 4300) 1.25v LLC3 as found by 1usmus' CTR tool
encoded 1806 frames in 190.61s (9.47 fps), 11820.04 kb/s, Avg QP:29.09

Just curious, what is the effective TDP between these two? 1usmus' CTR tool seems to be great at getting more efficiency, so I wonder if that shows here too. Thanks in advance!

.vodka · Oct 7, 2020

moinmoin said:
Just curious, what is the effective TDP between these two? 1usmus' CTR tool seems to be great at getting more efficiency, so I wonder if that shows here too. Thanks in advance!

Have a look.

Tweaked boost, -0.075mV undervolt, stock PPT TDC EDC (142 95 140) LLC auto

Per CCX (4375, 4375, 4325, 4300) 1.25v LLC3. This has no PPT TDC EDC limits, CPU will draw whatever it needs

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

Ignore most of the voltage readouts on the ASUS WMI section, the C6H has a crappy 22mV resolution steps for these things. Not reliable. The CPU's own SVI2 TFN telemetry on the other hand is reliable and the motherboard's VRM output agrees with the CPU's telemetry

Effective clock speed on every core is much faster on the manual overclock, power per core is lower, temperature is a few degrees higher, power consumption is similar although without any limits, it pulls more current. Increasing PPT TDC EDC limits on the boost algorithm doesn't result in what you see on the manual OC, just temperatures well into the 80°C range and lower clock speeds.

Per CCX OC is definitely more efficient if you have a good chip and a multithreaded enough workload to take advantage of it. Just need to be careful with vcore, ~1.25v is safe enough for the average Zen2 chip, but without knowing your chip's own FIT limit, it's a bit of an unknown on reliability and possible degradation if you hit it too hard with high current workloads most of the time.

I'm using a big air cooler. With a big AIO or custom loop, the boost algorithm could probably do better. Having said that, per CCX OC doesn't make much sense on this 3900x as it is a good sample that boosts high with these tweaks:

4375MHz results in ~540 single core, but the per CCX OC results in ~8900 MT (around a stock 3950x in this bench)

moinmoin · Oct 7, 2020

.vodka said:
Have a look.

Tweaked boost, -0.075mV undervolt, stock PPT TDC EDC (142 95 140) LLC auto

Per CCX (4375, 4375, 4325, 4300) 1.25v LLC3. This has no PPT TDC EDC limits, CPU will draw whatever it needs

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

Ignore most of the voltage readouts on the ASUS WMI section, the C6H has a crappy 22mV resolution steps for these things. Not reliable. The CPU's own SVI2 TFN telemetry on the other hand is reliable and the motherboard's VRM output agrees with the CPU's telemetry

Effective clock speed on every core is much faster on the manual overclock, power per core is lower, temperature is a few degrees higher, power consumption is similar although without any limits, it pulls more current. Increasing PPT TDC EDC limits on the boost algorithm doesn't result in what you see on the manual OC, just temperatures well into the 80°C range and lower clock speeds.

Per CCX OC is definitely more efficient if you have a good chip and a multithreaded enough workload to take advantage of it. Just need to be careful with vcore, ~1.25v is safe enough for the average Zen2 chip, but without knowing your chip's own FIT limit, it's a bit of an unknown on reliability and possible degradation if you hit it too hard with high current workloads most of the time.

I'm using a big air cooler. With a big AIO or custom loop, the boost algorithm could probably do better. Having said that, per CCX OC doesn't make much sense on this 3900x as it is a good sample that boosts high with these tweaks:

4375MHz results in ~540 single core, but the per CCX OC results in ~8900 MT (around a stock 3950x in this bench)

Wowzers, those are excellent numbers for MT indeed. Thanks again!

Hulk · Oct 7, 2020

Rank	User	CPU	GHz	seconds	fps/core/GHz	fps	Architecture	Cores
1	Fir	3990x 64/128	4.10	94.28	0.073	19.04	Zen 2	64
2	Fir	3970x 32/64	4.15	105.63	0.129	17.10	Zen 2	32
3	DaPoets	3960X 24/48	4.00	123.66	0.152	14.60	Zen 2	24
4	Charlie22911*	7980XE 18/36	4.37	141.12	0.163	12.80	Skylake	18
5	YBS1*	3950x 16/32	3.50	154.72	0.208	11.67	Zen 2	16
6	Micromd	2 - Xeon 8260L ES 48/96	2.80	156.99	0.086	11.50	Cascade Lake	48
7	Markfw	EPYC 7742 64/128	3.00	162.97	0.058	11.08	Zen	64
8	Bavor	3950x 16/32	4.05	169.73	0.164	10.64	Zen 2	16
9	Kjboughton	E5-2696 2x18 36/72	3.99	174.06	0.072	10.38	Broadwell	36
10	.vodka	3900x 12/24	4.35	190.61	0.182	9.47	Zen 2	12
11	DrMrLordX	3900x 12/24	4.35	191.78	0.180	9.42	Zen 2	12
12	rvborgh	Quad Opteron (48xK10)	3.00	200.00	0.063	9.03	Piledriver	48
13	Muadib	3900x 12/24	4.30	202.15	0.173	8.93	Zen 2	12
14	Cata40	7840xe 18/36	3.40	211.00	0.140	8.56	Skylake	18
15	Fir	2990WX 32/64	4.10	216.35	0.064	8.35	Zen+	32
16	Markfw	1950x 16/32	3.80	242.49	0.122	7.45	Zen	16
17	mrpiggy	E5-2699 18/36	3.70	253.41	0.107	7.13	Haswell	18
18	eek2121	1950x 16/32	4.10	254.04	0.108	7.11	Zen	16
19	RobertR1	9900k 8/16 (52C/47UC)	5.20	254.98	0.170	7.08	Coffee Lake	8
20	Lexingtonian	Quad Opteron 6386	3.50	275.66	0.029	6.55	Piledriver	64
21	Kenmitch*	Ryzen 7 3700x 8/16	3.97	279.43	0.203	6.46	Zen 2	8
22	CricketSP	9900k 8/16	4.90	281.80	0.163	6.41	Coffee Lake	8
23	kmm0000	9900k 8/16	4.60	290.96	0.169	6.21	Coffee Lake	8
24	Makaveli	3800x 8/16	4.90	292.40	0.158	6.18	Zen 2	8
25	TillausHamburg	Xeon ES QL1L 20/40	2.00	297.54	0.152	6.07	Skylake SP	20
26	sdifox	3700x 8/16	4.00	306.52	0.184	5.89	Zen 2	8
27	Chaotic42	9700k 8/8	5.00	311.03	0.145	5.81	Coffee Lake	8
28	XabanakFanatik	5960x 8/16	4.70	319.84	0.150	5.65	Haswell	8
29	DooKey	7820X 8/16	4.50	320.75	0.156	5.63	Skylake	8
30	epsilon 84	8700k 6/12 (5/4.9)	5.00	340.42	0.177	5.31	Coffee Lake	6
31	JoeRambo	8700k 6/12 (4.8/4.4)	4.80	342.38	0.183	5.27	Coffee Lake	6
32	mjdupuis	8700k 6/12 (4.8/4.8)	4.80	347.50	0.180	5.20	Coffee Lake	6
33	IEC	8700k 6/12 (4.7/4.7)	4.70	359.41	0.178	5.02	Coffee Lake	6
34	HairNuts	8700k 6/12	4.90	360.37	0.170	5.01	Coffee Lake	6
35	aclos3	3600x 6/12	4.04	367.20	0.203	4.92	Zen 2	6
36	CakeMonster	8700k 8/12 (4.9/4.6)	4.90	379.09	0.122	4.76	Coffee Lake	8
37	cmdrdredd	8700k 6/12 (4.7/4.5)	4.70	386.65	0.124	4.67	Coffee Lake	8
38	rgba	8700k 6/12	4.30	387.57	0.135	4.66	Coffee Lake	8
39	Markfw	E5-2696v3 14/28	2.50	394.29	0.131	4.58	Broadwell	14
40	vernieman	Intel 10600k 6/12	4.80	398.30	0.157	4.53	Comet Lake	6
41	Yeroon	Ryzen 3600 6/12	3.60	402.25	0.208	4.49	Zen 2	6
42	Markfw	2700x 8/16	4.10	404.44	0.136	4.47	Zen+	8
43	Dan057	Ryzen 2700 8/16	4.00	410.08	0.138	4.40	Zen+	8
44	lightmanek	Ryzen 7 2700x 8/16	4.00	414.34	0.136	4.36	Zen+	8
45	.vodka	Ryzen 1700x 8/16	3.95	416.04	0.137	4.34	Zen	8
46	pjmssn	Dual Xeon E5-2690 16/32	2.90	416.47	0.093	4.34	Sandy Bridge EP	16
47	Bburninatortech4	2700x 8/16	4.00	423.90	0.133	4.26	Zen+	8
48	Franz316	2700x 8/16	3.80	429.95	0.138	4.20	Zen+	8
49	AnnoyedGrunt	Ryzen 7 2700x 8/16	4.10	430.94	0.128	4.19	Zen+	8
50	richaron	Ryzen 1700 6/12 (Ubuntu 18.04)	3.70	434.68	0.187	4.15	Zen	6
51	CrackRabbit	2700x 8/16	3.80	440.29	0.135	4.10	Zen+	8
52	Despoiler	Ryzen 1700 8/16	3.90	444.03	0.130	4.07	Zen	8
53	ArisVir*	2700 8/16	3.80	454.14	0.131	3.98	Zen+	8
54	dfenton	Ryzen 1700 8/16	3.70	476.90	0.128	3.79	Zen	8
55	CuriousMike	9600k 6/6	4.40	477.30	0.143	3.78	Coffee Lake	6
56	Xenon	Ryzen 5 1600AF 6/12	4.00	504.29	0.149	3.58	Zen+	6
57	monkeydelmagico*	Ryzen 1600 AF 6/12	4.09	504.45	0.146	3.58	Zen	6
58	William Gaatjes	Ryzen 2600 6/12 (memory 3200)	3.60	519.91	0.161	3.47	Zen	6
59	mjdupuis	Ryzen 2600x 6/12	4.20	520.30	0.138	3.47	Zen+	6
60	ChippyUK	i5-8400 6/6	4.00	558.91	0.135	3.23	Coffee Lake	6
61	ywasmd	i5 8400 6/6	4.00	562.68	0.134	3.21	Coffee Lake	6
62	gamervivek	Ryzen 1600 6/12	3.50	593.10	0.145	3.05	Zen	6
63	CronoReverse	4790k 4/8	4.50	644.52	0.156	2.80	Haswell	4
64	Hulk*	4770K 4/8	3.90	754.96	0.153	2.39	Haswell	4
65	jones377	4770K 4/8	3.90	768.63	0.151	2.35	Haswell	4
66	Nathan Gall	4770K 4/8	3.90	775.24	0.149	2.33	Haswell	4
67	Junky228	Xeon w3690 6/12	4.00	877.63	0.086	2.06	Westmere EP	6
68	Imported_stryfe	2600k 4/8	4.50	951.72	0.105	1.90	Sandy Bridge	4
69	Bubbleawsome	i5-7300HQ (laptop) 4/4	3.20	1022.42	0.138	1.77	Kaby Lake	4
70	Hulk*	8250u 4/8 (Surface 2 laptop)	2.50	1067.13	0.169	1.69	Kaby Lake R	4
71	NickatNight1980	FX8320 4/8	4.16	1117.63	0.097	1.62	Piledriver	4
72	Xenon	i5-1035G4 4/8	2.10	1126.01	0.191	1.60	Ice Lake	4
73	Xenon	i5-3570k 4/4	4.00	1206.06	0.094	1.50	Ivy Bridge	4
74	Xenon*	FX-6300 6/6	4.10	1336.18	0.055	1.35	Piledriver	6
75	mylesrom	i5-3570 4/4	3.80	1399.56	0.085	1.29	Ivy Bridge	4
76	dfenton	i5-3570s 4/4	3.60	1586.81	0.079	1.14	Ivy Bridge	4
77	Imported_stryfe	i5-750 4/4	2.80	2012.80	0.080	0.90	Lynnfield	4
77	Slim Fan	Pentium Silver J5005 (NUK) 4/4	2.80	2631.13	0.061	0.69	Goldmont Plus	4
78	PeterScott	Core 2 Quad Q9400 4/4	3.20	2655.10	0.053	0.68	Yorkfield	4
79	ywasmd	i5 3210 (Macbook Pro) 2/4	3.10	3144.79	0.093	0.57	Ivy Bridge	2
80	Fanatical Meat	Q9650 4/4	3.00	3439.20	0.044	0.53	Yorkfield	4
81	dfenton	Celeron N3450 (laptop) 4/4	2.20	5099.43	0.040	0.35	Apollo Lake	4
82	dfenton	Celeron 2957U 2/2	1.40	6027.14	0.107	0.30	Haswell	2

vernieman · Oct 8, 2020

Just did this:
HandBrake 1.3.3 (2020061300)
OS: Microsoft Windows NT 10.0.18363.0
CPU: AMD Ryzen 3 3300X 4-Core Processor
RAM: 16GB DDR4-3200 T-Force
Mobo: MSI MPG B550I Gaming Edge Wi-Fi
GPU Information: AMD Radeon RX 5600 XT - 26.20.15029.27017
Cooling: MAG Core Liquid 240R
Storage: Samsung 960 EVO 250GB, SanDisk Plus III SSD 480GB, WD Blue SN550 1TB, WD Red 6TB
PSU: EVGA 500W 80+ Bronze
Screen: 1920x1080

encoded 1806 frames in 533.20s (3.39 fps), 11820.04 kb/s, Avg QP:29.09

Hulk · Dec 6, 2020

It would be nice to get a couple Zen 3 and Comet Lake scores on the board.

Justinus · Dec 6, 2020

Hulk said:
It would be nice to get a couple Zen 3 and Comet Lake scores on the board.

When my motherboard shows up in a few days I'll post 5950x scores.

bgt · Dec 6, 2020

HandBrake 1.3.2 (2020050300)
OS: Microsoft Windows NT 10.0.19042.0
CPU: AMD Ryzen 5 3600X 6-Core Processor
Ram: 49077 MB,
GPU Information:
NVIDIA GeForce GTX 1050 - 26.21.14.3200
Screen: 3840x2160

CPU is not OCed
-------------------------------------------

encoded 1806 frames in 373.32s (4.84 fps), 11820.04 kb/s, Avg QP:29.09

EvilRick · Dec 6, 2020

HandBrake 1.3.3 (2020061300)
OS: Microsoft Windows NT 10.0.19042.0
CPU: AMD Ryzen 7 2700 Eight-Core Processor
Ram: 16334 MB,
GPU Information:
AMD Radeon RX 5700 XT - 27.20.14501.18003
Screen: 3440x1440

(4.1 GHz | VCORE = 1.35 | SoC = 1.0125 | LLC =3 | DRAM = 1.4)

encoded 1806 frames in 394.30s (4.58 fps), 11820.04 kb/s, Avg QP:29.09

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

Senior member

Lifer

Member

Diamond Member

Senior member

Member

Golden Member

Senior member

Lifer

Moderator Emeritus, Elite Member

Member

Senior member

Junior Member

Junior Member

Junior Member

Golden Member

Diamond Member

Golden Member

Diamond Member

Diamond Member

Junior Member

Diamond Member

Diamond Member

Senior member

Member