Overclock.net banner
1 - 20 of 96 Posts

·
Registered
5900X, RTX 3080
Joined
·
854 Posts
Discussion Starter · #1 · (Edited)
What Is It?
4K x265 (HEVC) Benchmark
is an easy to use benchmark created by me for the OCN community (I decided to create this after it was suggested that Handbrake should be used to benchmark, which is not only far more complicated that it should be, but it also advocates for terrible software).

This benchmark utilises AVX (AVX, AVX2, and AVX512 if your CPU supports it) and is simple to run. It takes approximately 6 minutes to complete on a 12C/24T CPU. An included 10-bit 4K video file (Creative Commons) is encoded to 5,000 frames (03m 28s). If all 5,000 frames are encoded, a line of results is produced containing the following information:
  • If the benchmark completed without errors ("SUCCESS")
  • How long the benchmark took to complete (in seconds)
  • How many frames were encoded per second (fps)
  • The data/bit rate for encoding (kb/s)
  • Average QP (not yet relevant, but will be in future versions)
  • The date and time in local format (e.g. 🇬🇧 & 🇪🇺 = dd/mm/yyyy with a 24h clock, 🇺🇸 = mm/dd/yyyy with a 12 hour clock)
  • CPU details, including the brand and model (this will be changed in future version to include cores and logical processors (threads))
If all 5,000 frames are not encoded, an error is displayed with more information and suggestions to try and fix the issue.

The current x265 encoder version is 3.5+21-3898a4996 (Jan. 2022 compile)

Does it Work?
Yes.
I've done as much testing as is possible and this is being actively developed (it's taken about 15 hours to get to a releasable state). If you find any problems, please let me know.

How Long Does it Take?
That entirely depends upon your CPU, whether it's overclocked, RAM clocks, and other factors. Here are a few examples showing approximate benchmark times on selected CPUs:
  • AMD Ryzen 9 5950X: ~3m 45s
  • Intel Core i9-12900K: ~4m 30s
  • AMD Ryzen 9 5900X: ~6m
  • AMD Ryzen 7 5800X: ~7m
  • Intel Core-i9 9900K: ~8m
  • Intel Core-i5 1035G4: 31m (Surface 7 Pro)
Do I Need to Install Anything?
No.
The benchmark is entirely portable. You simply need to open the "Start Benchmark" file, read the short piece of information displayed, and then press any button to start the benchmark.

Is The Benchmark Safe?
Yes.
Here are the two primary files used in the benchmark. They have been scanned with VirusTotal and you can see the results by clicking on the link.
Where Do I Download It?
Size: ~274MB
Download (pCloud):
4K_x265_HEVC_Benchmark-v0.1.2.zip
SHA256 (pCloud): 4K_x265_HEVC_Benchmark-v0.1.2.zip.md5

Posting Results
Please post:
  • A screenshot of your score
  • If your CPU is overclocked (e.g. "Intel 12900K @ 5.2 GHz")
  • Your RAM clocks and generation (e.g. DDR4 3600 / DDR5 5200 )
Example Results / Sreenshots
This is purely a demonstration as I had about 90 tabs open and 30 explorer processes running when running this benchmark. I did it to take the screenshots to post here. I will post my actual score later on.
Font Screenshot Electronic device Computer Software


This is the screenshot that should be posted. It displays the fps, kbps, and CPU model.
Computer Personal computer Font Screenshot Software


Feedback is welcome.
 

·
Still kinda lost
Joined
·
3,697 Posts
Hey, ran it twice and both runs exited the second it finished. so if you can make it spit a log before exit that would be nice.
Was able to catch the second run filming with my phone.
Thanks for the effort

Running an Acer Nitro 7, i7 10750h, 16gb 2933mhz cpu was running between 48watts and 65watts (Other models can go upto 107watts, not a much bigger difference in most scenarios as these chips are designed to run perfectly at 45w-55w but it can get upto 10% in the right loud) it took 823.66 seconds for me.

Noticed a couple of thing as I use AVIDEMUX and HandBrake on weekly bases, CPU utilization was a little weird as it will idle for a 1 second every 1 to 2 mins and it will lower the usage to 65%-75% for couple seconds every now and then. considering wattage was fluctuating too means it's not forcing a 100% exactly like HandBrake.

But this might be a usage problem with either my specific machine or laptops in general, considering everything this is not far off from a realistic HandBrake load.

EDIT: HandBrake was written as hand break for some reason, I need to go to sleep.
 
  • Rep+
Reactions: Piers

·
Registered
Joined
·
675 Posts
Are you guys really sure you are running AVX2 @ 5.3 and 5.5 ghz ?
AVX2 instruction set is already limited by Intel to a multiplier of x51. While the set clock rate of 5.2 GHz, for example, is continuously available in applications without AVX2, such as Cinebench, and benchmark high scores can be achieved, the picture looks different in applications like LinpackXtreme. Even if you set the CPU clock to static in the BIOS, remove all performance limits and ensure sufficient cooling, the CPU throttles itself down to 5.1 GHz when executing AVX2 instructions. It also doesn’t matter how warm the CPU runs, how much power it consumes, or what the application is called. In HWInfo the intervention of the clock limit is recognizable by “IA: Max Turbo Limit – Yes”.

Rectangle Font Parallel Screenshot Pattern



Why Intel limits the clock speed for AVX2 is not confirmed yet. One possible reason could be fear of degradation due to electron migration at too high current levels – a phenomenon where the CPU slowly damages itself over time. To be fair, due to the generally very high heat density of the new CPUs, only very few users probably use an appropriate cooling system to ever be able to run such high clock rates with AVX2. But for example who uses a custom water cooling loop or even delids the CPU in search of the absolute performance maximum, will sooner or later come up against this artificial limit of these otherwise “unlocked” CPUs. By the way, this limit was already present in Rocket Lake CPUs and was probably just not really covered because the product generation was generally much less interesting.
Fortunately, there are already workarounds for both of these AVX hurdles, the throttling of AVX2 and the removal of AVX-512. For example, Asus has implemented a patch in their BIOS versions for “Maximus” series motherboards that disables AVX2 throttling. The only important thing here is that the clock must already be set in the BIOS at boot time. A subsequent change via in-OS software will otherwise get caught in Intel’s catch net again.

Check with:
Rectangle Font Parallel Magenta Circle
 

·
Registered
5900X, RTX 3080
Joined
·
854 Posts
Discussion Starter · #12 ·
Hey, ran it twice and both runs exited the second it finished.
Bizarre. Does Event Viewer record anything?
if you can make it spit a log before exit that would be nice
That's functionality I plan to add. I realise it's kind of essential.
Noticed a couple of thing as I use AVIDEMUX and HandBrake on weekly bases, CPU utilization was a little weird as it will idle for a 1 second every 1 to 2 mins
Both use FFmpeg, but much older version. The build I use for this was only compiled 3 days ago with the latest pull requests. I wouldn't focus on power use with x264 or x265 as it's perfectly normal to see wild fluctuations depending on whether the CPU is waiting for another component (e.g. storage). Remember, this is a 4K 10-bit x265 source file that's been encoded at --slower and is generally more complex to read.
and it will lower the usage to 65%-75% for couple seconds every now and then. considering wattage was fluctuating too means it's not forcing a 100% exactly like HandBrake
That's also perfectly normal for x264 and x265 and of course is source-dependent as well. HandBrake may have different optimisations or, far more likely, is using more power due to built-in filters and its odd way of handling frame rates such as 23.976 (broken since it was first released and uses workarounds as a 'fix').

PS: avidemux is a great piece of software. Handbrake is truly awful. Please consider a much more powerful GUI such as Staxrip. It's FOSS and offers so much more flexibility.
 

·
Registered
Joined
·
129 Posts
Are you guys really sure you are running AVX2 @ 5.3 and 5.5 ghz ?

Fortunately, there are already workarounds for both of these AVX hurdles, the throttling of AVX2 and the removal of AVX-512. For example, Asus has implemented a patch in their BIOS versions for “Maximus” series motherboards that disables AVX2 throttling. The only important thing here is that the clock must already be set in the BIOS at boot time. A subsequent change via in-OS software will otherwise get caught in Intel’s catch net again.

Check with: View attachment 2543802
Yes i have 5300 with AVX 2.0 litl bit improve witch 6400t1 memory.
Font Line Screenshot Electric blue Parallel
 

·
Banned
Joined
·
4,672 Posts
Why my 5300 fast then you 5500?
Gear 1 D4 vs Gear 2 D5. Yours has the edge in number crunching. Mine is more of a gaming computer. My Firestrike/TS/PR/SoTRwould likely be much higher.

AVX-512 is also disabled on mine. If I roll back to an old microcode the score could improve.
 

·
Registered
Joined
·
129 Posts
Gear 1 D4 vs Gear 2 D5. Yours has the edge in number crunching. Mine is more of a gaming computer. My Firestrike/TS/PR/SoTRwould likely be much higher.

AVX-512 is also disabled on mine. If I roll back to an old microcode the score could improve.
I have game computer too lol. SOTR 365 FPS. But i can improve if overclock more. I have new microcode too lol im on new bios your ddr4 lose fps on all game my ddr5. Lol and i Win Time spy etc. you have bad computer
 

·
Registered
Joined
·
3,727 Posts
Computer Personal computer Font Electronic device Multimedia
 
  • Rep+
Reactions: domdtxdissar

·
Banned
Joined
·
4,672 Posts
I have game computer too lol. SOTR 365 FPS. But i can improve if overclock more. I have new microcode too lol im on new bios your ddr4 lose fps on all game my ddr5. Lol and i Win Time spy etc. you have bad computer
You modified your level of detail and your graphics, physics, and combined scores are STILL lower. Doesn’t sound much like a gaming computer if you have to turn down the detail to get acceptable performance.

SU-35

F-22
 
1 - 20 of 96 Posts
Top