• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

3dMark Feature test, when CPU and API count.

sblantipodi

sblantipodi

2[H]4U
2FA
Joined
Aug 29, 2010
Messages
3,805
This is my results with the system in signature.

i7 5930K six cores CPU @ 4.2GHz
GTX980 SLI

http://www.3dmark.com/3dm/6427158

Post your results.

3dMark_featuretest.jpg
 
seems pointless to even run it with out direct x12.....but i can show my difference anyway
Capture_zpsxtkaa7u8.png
[/URL][/IMG]
 
skline00 said:
Here's mine:
http://www.3dmark.com/aot/9991

2v91wsk.jpg
Click to expand...

interestingly my old 7970 and ancient cpu is about 12 times more performance using dx12 over your dx11 but less than half the power of your (2) 290x's on mantle..... you probably triple my score if you run it on win10 and WDDM2 driver.....shows a lot of promise for WDDM2 games once there out if you think about it...to bad we still have a long wait for actual games

Admittedly the mantel performance on your is nice as is...to bad its not really ever been used to the fullest
 
primetime said:
interestingly my old 7970 and ancient cpu is about 12 times more performance using dx12 over your dx11 but less than half the power of your (2) 290x's on mantle..... you probably triple my score if you run it on win10 and WDDM2 driver.....shows a lot of promise for WDDM2 games once there out if you think about it...to bad we still have a long wait for actual games

Admittedly the mantel performance on your is nice as is...to bad its not really ever been used to the fullest
Click to expand...

GPU can't be compared using that benchmark, even futuremark say:

original.jpg


its a API Overhead Performance it have nothing to do with the raw gpu performance its all about CPU draw calls.. sadly i don't have the paid edition of 3dMark to test.. T_T
 
I see it as a test to see if your CPU can actually talk to the GPU to send it commands. Seems that most games have only been able to use one or two threads to send information to the GPU. DX12 is supposed to remedy that. Here is an AMD Blog on it. I'd assume that the AMD FX-8370 would be seeing the same utilization as the big boy Intel chips in this thread. I also wish I had a copy of the test suite to compare for fun.

http://community.amd.com/community/...mance-in-new-3dmark-api-overhead-feature-test
 
sblantipodi said:
It seems that AMD isn't able to scale well on multi thread without mantle
Click to expand...

Bad thing is that it's basically the same with Intel from what I hear. DX11 has been choking the life out anything that uses more than 2 cores. Luckily Intel has been working on better single core performance over the years so there was a small but measurable upgrade path. DX12 alleviates it by actually scaling up to 6 cores at least according to the AMD blog post.

Wonder how your 5930X scales with 2, 4, 6, virtual cores enabled / off, etc under DX11 and DX12. Tempted to buy 3DMark, but I can't wrap my brain around $25 to run a test once. Should have bought it for $5 when it was on sale in December. :(
 
cageymaru said:
Tempted to buy 3DMark, but I can't wrap my brain around $25 to run a test once. Should have bought it for $5 when it was on sale in December. :(
Click to expand...

Im in the exactly same boat.. right now was in the 3dmark page tempted to buy... :p
 
lol now would be a good time for it to go on sale again right? they would make the money with all the people curious right now.not sure if i paid 5 or 10 but i know i didn't pay full price
 
Araxie said:
GPU can't be compared using that benchmark, even futuremark say:

original.jpg


its a API Overhead Performance it have nothing to do with the raw gpu performance its all about CPU draw calls.. sadly i don't have the paid edition of 3dMark to test.. T_T
Click to expand...

while i guess i misspoke if it was taken as gpu performance...as i am aware its a combination of performance between the gpu and cpu with draw calls alone with different API's but i am seeing large differences between different cards being used and in some cases much weaker hardware doing better in the dx12 area then much newer and normally hugely more powerful hardware. Of course newer builds of win and newer driver from both teams will change results dramatically in the future but it is very interesting to see things develop. Now if i got this incorrectly again please feel free to correct me again, since im here to learn and eventually get a grasp of whats going on:D

Now would be a nice time for star swarm to release an update for us to play with instead of only letting the review sites have access to it.;) And in case any one missed it , 10049 is out to be are test version now and hopefully more new drivers.
 
Last edited:
cageymaru said:
Bad thing is that it's basically the same with Intel from what I hear. DX11 has been choking the life out anything that uses more than 2 cores. Luckily Intel has been working on better single core performance over the years so there was a small but measurable upgrade path. DX12 alleviates it by actually scaling up to 6 cores at least according to the AMD blog post.
Click to expand...

DX and OpenGL are, with few exceptions, serial APIs, where the rendering pipeline is done in steps. That's why you saw games not scale well beyond two cores; one core was handling the main game executable, and another was running the primary render thread. Those two threads handled about 90% of the total workload, and thus drove performance.

DX11 introduced a preliminary form of multithreaded rendering, but there were very strict requirements on usage. Games running Crytek 3 and Frostbite 3 do scale better as a result though, as the engines are coded to take advantage of this whenever possible.

DX12 is doing two things. The first is greatly reducing the overhead of the driver layer, which is basically going to give an across the board performance increase. The second is the ability to more easily do multithreaded rendering, which should allow workload to be broken up into smaller units, which should scale better across CPU cores. Do note: This will NOT automatically give a performance increase, unless the CPU in question was being overworked under DX11.

I actually expect the core i3 line to see the largest performance gains overall, since the i3's were the most bottlenecked CPUs out there, and thus benefit most from DX12 lowing CPU side requirements. I'd go so far as to argue that i3s will become legitimate gaming CPUs when DX12 starts getting used.
 
You must log in or register to reply here.
Back
Top