The results in the benchmarks v.5
In this topic, measured by PI... parrots in popular benchmarks.Previous topic
Note: screenshots must be hidden under the spoiler! Otherwise, posts that contain screenshots, not removed under the spoiler, will be deleted!
Vinni-Pukh
Damn, just sadistic and all, I wonder what it feels like
launched the game, ran in the beginning, where the horse is given)
At maximum incl. volocom gives 60-75 fps. Never drops below 60. C off. volocom fps 75-95. But this, again, in the initial location after training... What will happen next I do not know the game off, so as not to spoil the experience. First, it is necessary to pass the second part)
PS: forgot to mention that the cards are not yet persecuted, played on the head. frequencies (1316/7012)
-Sergey Viktorovich-
At maximum incl. volocom gives 60-75 fps.
And just for Garolite or at all?
Vinni-Pukh
And just for Garolite or at all?
the greatest Bank at all. I say, put everything on max that is possible.
PS: YouTube is full of clips where sly 970 from playing Witcher, test. But not all settings are shown)
In General, these days have tried benci 3дмарков any and havenow on 10 wines, wines 8.1 and win 7. And here are the results:
Win 8.1:
3dmark 11 - X8994
3dmark 11 - P17736
3DMark FS Exrtreme - 8993
Heaven 4.0 - 108.3 (2727)
Valley 1.0 - 106.9 (4474)
Win 7:
3dmark 11 - X9368
3dmark 11 - P18600
3DMark FS Exrtreme - 8427
Heaven 4.0 - 108.5 (2734)
Valley 1.0 - 107.3 (4488)
On 10-TKE results like the 8.1.
In 3дмарках a striking difference. If in the old Marche of the results on the 7-ke is much higher in the new - just the opposite. Heweny/valley - without a difference. The most stable bench, independent of the OS.
PS: graphics card without overclocking
-Sergey Viktorovich-
Interesting of course you guys I'm amazed.
Straight took was zabotilis into Windows and ran the benchmark straight I sat down and went as they say.
It doesn't quite work)
-Sergey Viktorovich-
I understand you filmed the results first to test the card and performance on different axes. The results (well, as far as I can judge from the figures) overall, not only graphics, therefore, you get the expected result: the haven no CPU score is the result of all smooth, 3Dmark'Ah is, CPU (physics) score - the result is different.
Of all the used tests, only the FSE was released and patched after the release of 8.1, therefore only in him skalitsa graphics from DX11.2, which consequently makes it faster in Win8.1
Now on the topic of the 11th:
It is obvious that in 11 of the Marche in the presence of 2 (not even overclocked cards) PERC (more physics score) will be the bottleneck. And as you know, the smallest of under results has the greatest impact on the overall result, that is, for example if you have 15K on physics and 30K on the chart, an extra 1K on the schedule will give you at best 150 points in the end result, and increase the thousands of the physics will give approximately fifteen thousand (i.e. 10 times more) parrots. Accordingly, the difference you got between the different axes caused largely by the difference in physics than graphics.
Due to the nature of work of the kernel of Windows (or rather the subtleties of how the wind carries the threads across the cores) on Win7 11 mark for physics is always faster than Win8. An empirical fact.
11th mark, in contrast to other Markov sensitive enough to twicam in fiziks test. This is probably the only mark which responds to tweaks memory (and cache of course). Other things being equal, one memory can elevate physics to the mower (or more if you have a good memory of the mother, and you know what to do with them).
Another small but significant caveat: in version 1.0.1 a result of the physics test is typically 8-10% more than the newer versions.
DartMaul
learned a lot of informative, thank you. I've never delved into such knowledge - tested in the brand after buying a new card, and well. No more)
Due to the nature of work of the kernel of Windows (or rather the subtleties of how the wind carries the threads across the cores) on Win7 11 mark for physics is always faster than Win8. An empirical fact.
but really.
3дмарк 11 x 8.1 wines
3дмарк 11 win 7
I especially never on the numbers of these individual did not pay attention, only the final result interesting.
in version 1.0.1 the result fiziks test typically 8-10% more than the newer versions.
I have 1.05 version
-Sergey Viktorovich-
I especially never on the numbers of these individual did not pay attention, only the final result interesting.
The end result without all the intermediate nothing says)
Is consistently,Above, have not tried,no time to test(((Maybe later, I put the haven
Spoiler.png)
.png)
.png)
Test in 3dmark Vantage
Settings not trogan.
OC win10 x64
Konfa:
i5 2400 1155 3.3 (Arctic mx2 thermal paste )
asusp8h61pro
8 gb hynix 1333
hd7950gb OC Sapphire at 1150 core and 1350 memory.
Max temperature of the CPU in the test, 51 degrees. 34-36 in just
Mac temperature GP 71 degrees idle 38
The Zalmun z11 Plus case (two 120mm on top, one behind and one in front, one 120 at the bottom. the block
Unit aerocool strike x 600w
I would like to know how the result on screen norms?
I can't find information how to test, what metrics to consider for a good system or bad
(p,s, assembled not a shop) he collected himself.
Spoiler
p.s. I think even the factory paste change on your Radik on MK2
Is 3dmark11
Spoiler
configuration• CPU: Intel Core i7 4790K @ 4.7 GHz (1.275 v, VID=1.070 v, not scalped)
• cooler: Arctic Cooling Freezer 240 Liquid
• MB: ASRock Z87 Extreme4
• RAM: 2 x 8GB Hynix 1600 MHz + 2 x 4Gb Hynix Orig. 1333 MHz @ 1600 MHz = 24 GB
• Video 1: MSI GeForce GTX 970 Gaming 100ME [1500/7800]
• Video 2: MSI GeForce GTX 970 Gaming 4G [1500/7800]
• OS: Windows 10 x64 Pro
3dmark 2011
3dmark 2013 v.2.0 Fire Strike Extreme
cinebench 11.5
cinebench 15
SVPmark 3.0.3 b