• 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 Time Spy DX12 test

I think this goes to show that pascal won't be as crippled as maxwell when it comes to the penalties of not being able to handle concurrent compute/graphics workloads. Must be the better preemption and ability to switch from a compute to a graphics workload, and vice versa that is allowing them to not be so handicapped.


This is why we need the bigger vega chips sooner, and higher clocked polaris chips. Fortunately, it looks like the latter will drop onto the market around the time the 1060 launches.
 
Async on vs off.

time_spy_results_async_vs_no_100671644_large.png
 
tybert7 said:
I think this goes to show that pascal won't be as crippled as maxwell when it comes to the penalties of not being able to handle concurrent compute/graphics workloads. Must be the better preemption and ability to switch from a compute to a graphics workload, and vice versa that is allowing them to not be so handicapped.


This is why we need the bigger vega chips sooner, and higher clocked polaris chips. Fortunately, it looks like the latter will drop onto the market around the time the 1060 launches.
Click to expand...

Pascal is also narrower (ALUs) and faster (clocks) so there is less of a dependency on concurrent work to fill the chip.
 
Either way this a win for everyone. Low level API will allow for cheaper cards to hit higher and higher performance gains simply by removing a lot of software overhead.
 
980 Ti is conspicuously missing, but I'm sure the difference between async compute on/off would be similar to GM104. But it would be nice to see how well it does with its handicapped async compute support compared to the Radeon cards with async compute.

This is PC World, though, so I wasn't expecting much before reading the article.
 
I'm concerned here where the 1060 will end up, is it 20% slower than the 1070? 25%?


if only 20% slower that will net it around 4450
if 25% slower that will generate a score of around 4225, Those are both higher than the base rx 480 in dx12.

amds current enemy is pascal not maxwell, we lost that battle with the nvidia fools blindly bought 970s and now they will suffer for it, but nvidia already has their money, so harping on that now does not help amd, now that they are going up against a less crippled pascal.
 
tybert7 said:
I'm concerned here where the 1060 will end up, is it 20% slower than the 1070? 25%?


if only 20% slower that will net it around 4450
if 25% slower that will generate a score of around 4225, Those are both higher than the base rx 480 in dx12.

amds current enemy is pascal not maxwell, we lost that battle with the nvidia fools blindly bought 970s and now they will suffer for it, but nvidia already has their money, so harping on that now does not help amd, now that they are going up against a less crippled pascal.
Click to expand...
The 970 was a great card. The fiasco over the VRAM was wrong on Nvidia for not being transparent, but it was also overblown as its hit to performance was negligable.
 
Presbytier said:
The 970 was a great card. The fiasco over the VRAM was wrong on Nvidia for not being transparent, but it was also overblown as its hit to performance was negligable.
Click to expand...

A great card chosen too often over a better performing card for the same money in the 390, that was always the point. perf/watt does not help gamers play games with better framerates and boosted performance, and people chose lower performance and a quicker upgrade requirement.
 
Presbytier said:
The 970 was a great card. The fiasco over the VRAM was wrong on Nvidia for not being transparent, but it was also overblown as its hit to performance was negligable.
Click to expand...

And the lies about the ROP count... and the L2 cache.... but yeah it was a great card when it came out.
 
When is this going to be released? Finally there moving away from firestrike!....i was hoping for more space battles like in the old days...oh wow. Cant wait to see how bad my card sucks with 3dmark dx12!:woot:
 
primetime said:
When is this going to be released? Finally there moving away from firestrike!....i was hoping for more space battles like in the old days...oh wow. Cant wait to see how bad my card sucks with 3dmark dx12!:woot:
Click to expand...

It's already released.
 
tybert7 said:
A great card chosen too often over a better performing card for the same money in the 390, that was always the point. perf/watt does not help gamers play games with better framerates and boosted performance, and people chose lower performance and a quicker upgrade requirement.
Click to expand...
The R9 390 was released 9 months after the GTX 970. Since the R9 390 came out I've seen more recommendations for it over the GTX 970, myself. The final purchase decision, though, is always up to the individual.
 
I'm running Win 10 Insider's Preview and my SSD is 90% full. R.I.P Hitachi RAID 0.

I scored 4 052 in Time Spy

3DMark Score
4052
Graphics Score
4145
CPU Score
3599
Graphics Test 1
27.83 fps
Graphics Test 2
23.18 fps
CPU Test
12.09 fps
 
F6lTM3M.jpg


Detailed benchmarks

9XvslAi.png
 
Last edited:
Main Gaming Rig
Asus Strix 1080 OC + 83 Core + 500 Memory
Graphic Score: 7813
CPU Score: 5093
Overall Score: 7233
Firestrike Graphic score 25328

Wife's Gaming Rig w/ 980ti
Zotac 980ti Amp Extreme + +100 Core + 500 Memory
Graphic score: 6266
CPU Score 4719
Overall Score: 5972
Firestrike Graphic Score 22537

Wife's Gaming Rig w/ 1070 OCed
MSI Gaming X 1070 + 125 Core +500 Memory
Graphic Score 6428
CPU Score 4594
Overall Score 6064
Firestrike Graphic Score 20983

Time Spy Asus Strix 1080 OCed vs Zotac 980ti Amp! Extreme OCed is about 24.7% Difference,
MSi Gaming X 1070 is 2.5% Higher than 980Ti but 21.5% Slower than 1080

Firestrike Asus Strix 1080 OC vs Zotac 980ti is only 12.4% Difference @ 1080p
MSI Gaming X 1070 is 7% Lower than the Zotac 980ti , 20.7% lower than 1080

so with the Async, Gaming X 1070 gained almost 9.5% boost over 980ti. and 1080 Gained another 12.3% over 980ti
 
Last edited:
With the paid upgrade, it lists the resolution as 2560x1440 when you go into the Custom settings.
 
3DMark Score
3200
Graphics Score
3022
CPU Score
4814

Not to bad considering I have an ancient 780, LOL.
 
primetime said:
Bah! No demo! lol it going to be a while before i buy it:(
Click to expand...
figured out that you can get the free version via the website but not steam. so I dl'd and ran and my poor old system gets a lowly 2514 :( free version is also 1440p.

3dmark.png
 
JustReason said:
In another forum overclockers.net Mahigan, you know you guys love him, gives a great explanation of async and Pascal.

[Various] Futuremark Releases 3DMark Time Spy DirectX 12 Benchmark - Page 22
Click to expand...

I guess this thing about all of that is so what at the high end? AMD has nothing that touches a 1080 right now. At the lower end for now AMD looks to have an advantage at least until next week when the 1060. And that's not to discount what AMD has here. But they still don't have a top performing product and no one really know how async carries over to things that don't exist currently.
 
heatlesssun said:
I guess this thing about all of that is so what at the high end? AMD has nothing that touches a 1080 right now. At the lower end for now AMD looks to have an advantage at least until next week when the 1060. And that's not to discount what AMD has here. But they still don't have a top performing product and no one really know how async carries over to things that don't exist currently.
Click to expand...
Well we kind of have to look at the big picture including a larger time frame than "now". AMD with DX12 and Vulkan really doesn't need to push higher tier cards, well rush them is likely the better terminology. In a lot of DX12(proper) titles AMD shifts their whole lineup up the performance graphs against where they were in DX11 graphs against Nvidia. Now for the foreseeable future I don't think Nvidia is in any kind of devastating trouble. However if Games start releasing DX12(proper) rather than DX11 then patched to DX12 and they start coming over from consoles then it gets a bit more concerning.

Honestly both AMD and Nvidia have a perfect setup as neither is competing with the other so sales are as good as they can get. Now that doesn't really help price on the high end(1080/1070) but does quell any price hikes on the low end(480).
 
MangoSeed said:
He's guessing just like everyone else. Just more eloquently.
Click to expand...
He's very obviously wrong, as he has been in the paste in his haste to point out some fatal flaw in nvidia's engineering :p

Dedicated an entire gpc to compute is a huge waste of geometry resources, it makes no sense at all. Preemption has nothing to do with async compute, I don't know why people keep mentioning it. Preemption is almost antithetical to async compute, because you don't want to preempt the task unless there's a stall
 
You must log in or register to reply here.
Back
Top