Cinebench R15 AMD results - new fun for next months

[FlanK3r]

wow my frined, wow....Can you add R11.5 score too?

[IronBalls]

Just got myself new ram. Crucial Elite DDR3 1866. They scale really well. Running stable @ DDR 2133 and Latencies 8 9 9 27 @ 1.6v. I can get them to work at 8 8 8 24 too, but I need a big voltage bump.

The point is, my R15 score is barely better... 854.



My best score with my old DDR3 @ 1600 and 9 9 9 24 was 849.

I don't understand how some of you get way better scores with the same CPU frequency ...

[FlanK3r]

try a bit lowet TRC

[IronBalls]

try a bit lowet TRC

I tried your advice. It does help a little bit.

I did an extensive overclocking session. Put my Mora-3 outside. Watertemp was initially @ 15C

Some funny pics of my setup





I achieved 859 in Cinebench R15.



My CPU doesn't scale with higher voltages. Anything above 1.575v will decrease overclockability. I even tried a run @ 1.625v and it crashed instantly.
Watertemps under 15C don't help either. Really strange CPU ...

I finally tried SuperPi @ 5450 Mhz and got 14.134s

This CPU is good for gaming, but not suited for hardcore overclocking.

[FlanK3r]

Lol, your clothes dryer seems very stable (big case at corner!). Impossible at my...chinese quality, light broken

But this CPU is good for 24/7 OC! 100%

[Daveburt714]

I don't understand how some of you get way better scores with the same CPU frequency ...

FWIW, CB15 seems to depend much more on CPU Freq than it does on IMC/Mem clocks...
A few Mhz on the core will normally result in better scores than trying to push IMC/Mem to the limit!
In fact, a lot of times pushing those clocks can actually have a negative effect on your final score (within reason OFC).

I'm probably giving away some secrets here, but once I find what my rig will run CB15 at, I switch to "Diagnostics Mode"/reboot and run it one last time for the best score without all the Windows crap running in the background!
(I don't feel this is cheating, just wise )

I must say, that's a golden 8320 you have there...

Get you best clocks and try running it using the above tips!

A good challenge may motivate me to push even harder...

[FlanK3r]

motivation could be this Dave... http://i.imgur.com/L8O0aeQ.png

If this chip can run these clocks at 22C ambients at air, what can doing in winter with open window? ,-)

[Daveburt714]

What little gem is that FlanK3r?

Voltages are kind of high, but those are some killer clocks for air cooling!
BTW, sorry to see you had problems with your AIO water, that sucks...
You always seem to get the best new stuff!!!

Will the NDA allow you to give us more info?
By the TDP I'm assuming it's a new 9xxx series chip (with core revisions )...

BTW: Get you water fixed/Beat my 871 and you'll definitely motivate me!

[FlanK3r]

Its old screen my friend, is simply FX-9590 in ES , So no need NDA for now :-D

[IronBalls]

motivation could be this Dave... http://i.imgur.com/L8O0aeQ.png

If this chip can run these clocks at 22C ambients at air, what can doing in winter with open window? ,-)

You cant know until you try and find for yourself. I can run Cinebench @ 5.4 with ambient temps but anything above seems impossible. Perhaps it will scale, or perhaps it will hit a wall @ 5.4 like mine. I also thought lower temps would allow me to bench a little higher but it wasn't the case.

But the CPU does look promising. You just have to cross the fingers it scales with lower temps.

[IronBalls]

FWIW, CB15 seems to depend much more on CPU Freq than it does on IMC/Mem clocks...
A few Mhz on the core will normally result in better scores than trying to push IMC/Mem to the limit!
In fact, a lot of times pushing those clocks can actually have a negative effect on your final score (within reason OFC).

I'm probably giving away some secrets here, but once I find what my rig will run CB15 at, I switch to "Diagnostics Mode"/reboot and run it one last time for the best score without all the Windows crap running in the background!
(I don't feel this is cheating, just wise )

I must say, that's a golden 8320 you have there...

Get you best clocks and try running it using the above tips!

A good challenge may motivate me to push even harder...

Thanks for the advices It never came to my mind to run Cinebench in Diagnostics Mode. It did come to my mind to use and old Hard drive, reinstall a clean copy of windows with only what is necessary. But your idea is easier to put in practice.

I did learn the hard way that this benchmark is cpu dependant, memory only improved my score a little. Nevertheless, I'm happy I bought the crucial elite kit, it does help in games.

I will probably retry someday, when temps outside are near zero, to run Cinebench in diagnostic mode.

[FlanK3r]

lol...many results are missing 2014 last backup?...

[DavidMMM]

a) first score: dual 6386 ES @ 4,2GHz 1,32V on a Supermicro H8DGI-F;
b) second and third scores: dual 6380 ES @ 4,1GHz on a Supermicro H8DGI-F;
c) fourth score: dual X5690 3,6GHz in turbo mode 2 x 6c/12t on a Dell T7500;

[I.nfraR.ed]

Let's bring new life to this thread


Ryzen 7 1700 @ 3.9GHz, 2x8GB G.Skill Trident-Z @ 1600MHz (DDR4-3200) 12-12-12-28 1T, Box cooler
1801 cb




Ryzen 7 1700 @ 4.11GHz, 2x8GB G.Skill Trident-Z @ 1600MHz (DDR4-3200) 12-12-12-28 1T, chilled water

1908cb




Ryzen 7 1800Х @ 4.33GHz, 2x8GB G.Skill Trident-Z @ 1613MHz (DDR4-3226) 12-12-12-28 1T
2002cb



Will upload images to XS later.

[chew*]

nice work looks like its scaling with memory.

[FlanK3r]

yeah, great score. Its scaling at 100% very good with memory clocks+timings. B-die, B-die, where are you?

Infra: How did you cooled it at 4333 MHz, its possbile get it with solid AIO and cold ambients (+5C in room)?

[chew*]

he is on chilled water. scales with cold not so much with volts.

[I.nfraR.ed]

Yes, they scale with volts till some point, then you need a big bump for very small increase in frequency.
Cold helps more in this case, but as chew* noted, I'm on chilled water.
Idle temp was about 5?C. Cold ambient should help, yes.
LN2 this weekend on the 1800X.

[Lightman]

Great numbers I.nfraR.ed

What volts were needed for these mem clocks and timings?

I need to sell my Haswell ASAP and get me some Ryzen!

[I.nfraR.ed]

What volts were needed for these mem clocks and timings?

1.45V just to be on the safe side, not sure how much I can lower it.

[charged3800z24]

Ryzen Stock Cooler, 1.48v, 4ghz

[DavidMMM]

Ryzen 7 1700 @ 4GHz, 1,36V, 2x4GB Patriot Viper 3000 - 1200MHz (DDR4-2400) 15-15-15-36 1T, FZ120 - 1733 cb

[FlanK3r]

Ryzen 7 1700X@5095 MHz 1.6x Vcore, LN2 cooling, score 2271 points.

[dave_graham]

here's an un-optimized Epyc 7601 Cinebench R15 score:
https://www.dropbox.com/s/dtv381bai6...53.02.png?dl=0

cheers,

D

[chew*]

1379 per die. Out of curiosity dave any chance you can run a model up or model down sku? Just need a variable clock speed wise for some simple math.

Puts a 2.2g threadripper around 2258.

I need to figure in the turbo variance albeit....probably clocked down i bet..