• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

FEATURED The 4X4 Summer Showdown! 2020

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.

Johan45

Benching Team Leader Super Moderator
Joined
Dec 19, 2012
OK, this summer is nearly a repeat of last summer's 4x4.
This time around I have selected 4 benchmarks that are to be run at 4.5 GHz on 4 cores with HT/SMT if you have it. We're also going to substitute Benchmate here for monitoring. Download the "Big" pack which has the CB and 7zip benchmarks within it. Run the benchmarks from the Benchmate console and hit save when finished. This will bring up a results window that is needed along with the CPUz tabs. HWBot x265 has it's own timer so it should show accurate results for the speed if I find it doesn't we'll have to add HWInfo. I have included a link for HWBot x265 below. So without further ado here we go:

What: 7ZIP, Cinebench R15, Cinebench R20, and HWBot x265 1080p
Who: any forum member
CPU: any CPU limited to 4 cores plus HT/SMT and 4.5 GHz can have more cores but must be disabled in BIOS
Cooling: Any
When: Ends: August 21, Midnight EST
Full Competition Rules: Here
As was stated a 4.5 GHz limit is in effect with a 10 MHz leaway (4510 MAX) as well as 4 cores. I ask that any submission that's put in uses Benchmate for monitoring speeds and must be open during the benchmark. CPUz tab must show only 4 cores being used so disable any extras in BIOS beforehand. The final score will be the total from all 4 benchmarks and please include this in your post. Also, use the supplied background

I would also like this format used at the top of your post and tally your scores for me please UserName/CPU/Score

BACKGROUND


Johan45/ R9 3900X/ H2O/ 52376.283
Johan45/ Ryzen 9 3900X/ H2O/ 52,232.283
dejo/ 3700x/ AIO360/ 52217.29
dejo/ Ryzen 7 3700x/ 360 AIO/ 52080.29
bassnut / 3900X / H20 / 50761.459
mllrkllr88 / 3900 OEM / AIO / 50394.718
jiccman1965/3800x/LN2/50,257
mackerel / 3700X / air / 49561.401
CADDI DADDI| 8086K| H20|40301.639
mackerel / 6700k / air / 41000.661
Caddi Daddi| intel i5, 4460| 20844.242
 
Last edited:
is this to say 4 cores and potentially 8 threads or just 4 threads?
 
4 cores and 8 threads

Another note, Benchmate and HWBot X265 require the HPET to be enabled. To do this open a command prompt as Admin and type bcdedit/set useplatformclock true then hit enter and reboot.
 
Last edited:
Here's an example
Johan45/ Ryzen 3 3100/ 360 AIO/ 43,284.106

7Zip 39896

7zip 39896.JPG

CB R20 2322

r20 2322.JPG

CB R15 1032

r15 1032.JPG

X265 34.106

x265 34.106.JPG
 
Last edited:
dejo/ Ryzen 7 3700x/ 360 AIO/ 52080.29
7zip 48051
7zip48051.jpg
CB R20 2756
20cine2756.jpg
CB R15 1232
15cine1232.jpg
Z264 41.29
x264hwbot41-29.jpg
 

Attachments

  • 20cine2766.jpg
    20cine2766.jpg
    1.1 MB · Views: 432
Last edited:
The way presented, can I assume each set of 4 results has to be from the same CPU model - no mix and match? I might have to submit sets from multiple CPUs as I'm not sure which ones are stronger in which benchmarks. Do know I wont have anything competitive AMD side.
 
The way presented, can I assume each set of 4 results has to be from the same CPU model - no mix and match? I might have to submit sets from multiple CPUs as I'm not sure which ones are stronger in which benchmarks. Do know I wont have anything competitive AMD side.

You'll be surprised Mack. When they're running the same speed AMD does very well against Intel and yes all subs from the same CPU

Hey Dejo you have CBR 15 up there twice
 
Last edited:
You'll be surprised Mack. When they're running the same speed AMD does very well against Intel and yes all subs from the same CPU

That's why I don't think my AMD CPUs will be competitive. I already know they wont hit the clock limit. They're really early retail samples, and not high end ones at that, and I've not been able to get all core much beyond 4.3 without a world of stability pain. Even limited to 4 cores, I'm not sure that'll buy me enough. If I can strap an Intel waterblock on it I could dust off the chiller but I'm not sure I'm motivated enough for that level. My best (easy) chance is possibly to get a high rank among Intel submissions.

That's not to say I wont revisit Ryzen, maybe the various bios updates have changed things since the early days, but it wont be a priority.
 
You'll be surprised Mack. When they're running the same speed AMD does very well against Intel and yes all subs from the same CPU

Some quick runs on [email protected] and all cores, and somehow my results are about as high or slightly better (literally single points) than yours but significantly worse than dejo's. Most of these benchmarks are actually better for AMD. Looking at the scores, AMD will probably beat Intel in some of these benchmarks even with 100-200MHz lower clock. For example, dejo has 48k in 7-Zip and I can barely get 37.5K.
 
Some quick runs on [email protected] and all cores, and somehow my results are about as high or slightly better (literally single points) than yours but significantly worse than dejo's. Most of these benchmarks are actually better for AMD. Looking at the scores, AMD will probably beat Intel in some of these benchmarks even with 100-200MHz lower clock. For example, dejo has 48k in 7-Zip and I can barely get 37.5K.

My run was at stock with the 3100 which is 3.9 all core
 
I haven't noticed ... now I'm sad :rain: Still the 10500 isn't bad in general. 6 cores and it runs fine using the Cryorig C7 cooler which is the size of Intel stock one. Quite a surprise considering that max temps are ~90°C and Ryzen 3600 would hit 95°C in like 2 mins on the same cooler.
 
I should note that I am rerunning the CB R20 test. I had issues with the screenshot tab on benchmate, it would often just pull up the last screen that was taken. Others may want to make sure what they are saving before assuming to correct one is there.
I will edit my post when I get time to get a new run in.
 
Most of these benchmarks are actually better for AMD. Looking at the scores, AMD will probably beat Intel in some of these benchmarks even with 100-200MHz lower clock.

I just bothered to look up my old IPC testing I did between Zen 2 and Skylake. I only have the chart to hand, not the raw numbers, but I was seeing over 10% advantage for Zen 2 over Skylake in CB20, and 20% in CB15. I didn't do a similar test for x265, and never ran 7zip bench so don't know how those will scale. Still hope to do Intel runs first, as I know they're an easy 4.5 GHz. Just dial it in and set voltage. Maybe tweak the ram a bit. Zen 2 will be a lot more work to get the best benching clocks as I'm not going to get 4.5 all cores.
 
dejo/ 3700x/ AIO360/ 52217.29
7zip 48143
7zip48143.jpg
CB R15 1245
15cine1245.jpg
CB R20 2788
20cine2788.jpg
X265 41.29
x264hwbot41-29.jpg
 
mackerel / 6700k / air / 41000.661

Ok, here's my initial sub to get on the boards and give myself some reference to work with. The Ryzen IPC advantage might take Intel out of the running here...

Not seeking to change the rules, but as an observation the 7zip scores are ball park 1000x that of x265, so it does impact the effective weighting of each bench significantly. If I want to improve my overall score, I'm going to spend a LOT more effort on 7zip than x265.

x265 1080p: 36.661
x265-1080-36-661.png

7zip: 37663

7-Zip_Launcher_1.1_37663.jpg

CB15: 987

CINEBENCH_R15_CPU_987.jpg

CB20: 2314

CINEBENCH_R20_CPU_2314.jpg
 
mackerel / 3700X / air / 49561.401

Had a go with the 3700X, and as suspected this sample/cooling wasn't going to do 4.5 without a fight, one I'm not sure I want to start yet. I settled on 4.4 (or a fraction lower due to the lower bus) and even then I had to use what I'd consider an uncomfortable voltage to get CB20 to not crash. Peak temps were still around 70C so if I'm feeling really brave there might be a bit more headroom to extract.

x265 39.401
x265-1080p-39-401.jpg

7zip 45682
7-Zip_Launcher_1.1_45682.jpg

CB15 1181
CINEBENCH_R15_CPU_1181.jpg

CB20 2659
CINEBENCH_R20_CPU_2659.jpg
 
If we exclude CB15 then your 6700K results are pretty much as high as my 10500. At least in these benchmarks, results are pretty disappointing compared to Ryzen.
I have some B550 mobos to test so I guess I will back to AMD soon.
 
My Z77 is dead, but I have a new board coming! I don't want to run my X5690 anymore or else I would have posted. I've never actually disabled cores on this CPU, and I don't want to start now :D
 
Back