• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

AMD signs up 40 development studios for Mantle

In fairness it seems like the users on this forum rarely get the same numbers as review sites, so it seems like none of them can be trusted. We should probably just stop posting graphs from websites :)



Another good point. Microsoft's API will work for these people and may also improve gaming for people with laptops/desktops running Intel graphics. Hopefully DX 12 will too, although I believe it still remains to be seen what OS is needed for DX12. These users may not bother with a new OS until they buy a new system.
Microsoft is helping these people to game on their PCs, even if it could be done better. There are probably people today that will play games on these PCs thanks to Microsoft. What Microsoft and DirectX has done for a long time is open up gaming to a large number of users.
There are also a large number of PC users, including AMD users that probably don't ever update their graphics drivers and so won't get Mantle, they'll have DirectX though.
There are also a large number of people with older AMD graphics cards that don't support GCN and DirectX allows them to play games.

Linux will struggle to get these average users I think. I think it'll struggle to get 'super users' too until it has a good backlog of games. How many people here still go back to Skyrim, WoW or some other DirectX game from time-to-time? Would you give this up so you could play the 2 new OpenGL games and Valve's back-catalogue? And then the trouble is, which devs will go to the effort of making a game run on an OS nobody is really using? So how will it ever get a decent back-catalogue? Why does it matter to devs if we use Windows or Linux?

Mantle will only be of benefit to newer AMD graphics card owners and also only those that update their drivers or had Mantle drivers pre-installed. And even then unless it becomes the default graphics option there are normal users that may not use it.

The 'average' user/gamer is likely to owe their gaming experience more to Microsoft than Linux or AMD. If this is Microsoft hating on PC gamers what must these other companies that are doing less for gamers think of PC gamers?

Could Microsoft have released a new major version of DirectX sooner than they have (DX11.2 was fairly recent as I recall)? Yeah, probably. But their are probably lots of companies that don't update things as quickly as they could. I mean it took Microsoft, what, 10 years to update their precious console money cow that is the XBOX 360 to the XBOX ONE. So it looks like they (and Sony) let the console market stagnate as well if that's your definition.

Who has done more for PC gaming than Microsoft?
Could they have done better? Probably, very few things are perfect.
Could they have done worse/less? Yes, they could have done as much as everyone else has.

Anyone playing the latest and for future games will have updated there drivers that have support for the latest games regardless of Mantle.
 
Last edited:
If you read some of the help etc threads you might well be surprised how many dont have the latest drivers or dont run beta's. Thats on both sides btw not picking on Nvidia or AMD
 
Who has done more for PC gaming than Microsoft?
Could they have done better? Probably, very few things are perfect.
Could they have done worse/less? Yes, they could have done as much as everyone else has.
The problem with MS is that it has been giving the PC gaming community the cold shoulder for quite a while now, and it's difficult to take what they said about "we care about the PC community" seriously, with them sleeping with Xbone in one arm, and tablet and phone in the other :rolleyes:
 
If you read some of the help etc threads you might well be surprised how many dont have the latest drivers or dont run beta's. Thats on both sides btw not picking on Nvidia or AMD

I decided not to bother writing that part that when latest game/games does not run right for the users who dont update there drivers that they will ask for help and will be asked are you using the latest drivers thus they will end up using the latest drivers one way or another if they want to play the latest games, it does not matter if its DX,Mantle or OpenGL.
 
Last edited:
If you have a bench for me to replicate with your 337 Drivers, I'm happy to do it with my weak CPU in Mantle.

Replace the CustomScenario.csv in your SS folder with this: https://mega.co.nz/#!AYMGyABQ!19Mpus89txDgJ0Bmi39eevh6LMsUzs43_0lgQZSRiKY

And then run the bench with extreme settings using the custom scenario option. You should get a scenario with a fixed camera angle and fixed ship spawns (someone on here made the thing, can't remember who, but thanks to whoever it was).

It's completely different from the follow mode. With follow I've seen over 100fps on my rig.


Couldn't possibly have anything to do with CPU clocks, testing places within the game, etc.?

It's impossible to compare without knowing where and how these guys were testing.

Gamegpu.ru shows the same thing though with their testing of the latest BF4 expansion (at least it's the latest as far as I'm aware, I don't play the thing).

http://gamegpu.ru/action-/-fps-/-tps/battlefield-4-naval-strike-test-gpu.html

1080p avg/min:
780Ti: 97/74
290X M: 87/74
290X DX: 79/64

1600p avg/min:
780Ti: 60/43
290X M: 55/43
290X DX: 51/41

And they even have a youtube video of their testing routine so anyone can try it.

Of course I could always try and get BF4 working on my rig to try this stuff out myself.

However that's not really the point here, the point is that you can't universally apply a "terrible" tag on DX11 as humbug did earlier. AMD's and Nvidia's DX11 implementations are not equal at the moment. NV's is leaps and bounds above AMD's whenever you see a CPU bottleneck. Mantle might be ahead in efficiency but then again it's quite limited as far as the games library goes.
 
Going to try those 720p tests myself, because they definitely aren't right at all.

Really wish games like bf4 had a built in benchmark, so much easier that way.

I agree about the but in benchmark as there are too many variables in trying to do a benchmark on it, Standing still shows a bottleneck in one place for sure but playing the game normally looks to be much closer between mantle amd and dx nvidia, Not as close dx vs dx on both though which is very strange.
 
playing the game normally looks to be much closer between mantle amd and dx nvidia, Not as close dx vs dx on both though which is very strange.

It is, don't be put off by what reviewers say. Assuming the cards are of equal performance AMD is as fast, if not faster than Nvidia in BF4 on dx11.1. Me and PG Tips compared results. 4770K and 780 Water cooled SLI stock vs 2700k and 290P air cooled reference cards stock. Our results were virtually identical. I had slightly higher fps and slightly less variance and stutter but there was nothing in it. Both played on CQ Large 48 man Metro servers. ;)

I did manage some results, eventually

bf4frametimemetronew.jpg

LtMatt said:
290P Crossfire DX11.1 14.4 WHQL

yEG9oEN.jpg

That was at 1080P so prime area for AMD's supposedly bad DX performance.
 
Last edited:
Replace the CustomScenario.csv in your SS folder with this: https://mega.co.nz/#!AYMGyABQ!19Mpus89txDgJ0Bmi39eevh6LMsUzs43_0lgQZSRiKY

And then run the bench with extreme settings using the custom scenario option. You should get a scenario with a fixed camera angle and fixed ship spawns (someone on here made the thing, can't remember who, but thanks to whoever it was).

It's completely different from the follow mode. With follow I've seen over 100fps on my rig.



Couldn't possibly have anything to do with CPU clocks, testing places within the game, etc.?

It's impossible to compare without knowing where and how these guys were testing.

Gamegpu.ru shows the same thing though with their testing of the latest BF4 expansion (at least it's the latest as far as I'm aware, I don't play the thing).

http://gamegpu.ru/action-/-fps-/-tps/battlefield-4-naval-strike-test-gpu.html

1080p avg/min:
780Ti: 97/74
290X M: 87/74
290X DX: 79/64

1600p avg/min:
780Ti: 60/43
290X M: 55/43
290X DX: 51/41

And they even have a youtube video of their testing routine so anyone can try it.

Of course I could always try and get BF4 working on my rig to try this stuff out myself.

However that's not really the point here, the point is that you can't universally apply a "terrible" tag on DX11 as humbug did earlier. AMD's and Nvidia's DX11 implementations are not equal at the moment. NV's is leaps and bounds above AMD's whenever you see a CPU bottleneck. Mantle might be ahead in efficiency but then again it's quite limited as far as the games library goes.

No there not, they are about the same bar a few percent margin of error in DX11.

Anyway, something is wrong with your Custom File, it doesn't even work with Mantle.

 
Last edited:


seriously who wouldnt wanna se these improvements in all games, significant boost in all fps, but you gotta love the minimum fps going from 19 to 32, it even went 44 when i turned off automatic texture limit.
this bench was run on everything maxed on 1680 res, FX6300 ( stock frequency) and R9 280X vaporx (core 1075mhz).
if i didnt have a lousy bandwitch i would have steam downloaded Thief again to test on 8350.
personaly i think Mantle brings a lot for no cost at all, i dont understand why some ppl might be against it, it doesnt cost them anything, they still have DX, it's not like the games go Mantle and cancels Dx version, so other than being simply jealous, i dont see the reason to be against Mantle, and even it helped bring them Dx12, so instead of the hate they should be glad that AMD brought Mantle to the PC gaming veryone stands to win from it even Nvidia players.
 
Last edited:
Well i'll tell ya what _Alatar_, your custom Scenario is a stroll in the park, there is almost no strain on the system, least of all the CPU. its a child's play setting. all it does is remove the work the CPU needs to do in DirectX.
In other words it gets around the low performance batching problem in DirectX by reducing batching levels.

Whose setting is it? was it used in Mantle vs DirectX reviews?

===========================================================
Oxide Games
Star Swarm Stress Test - ©2013
C:\Users\#####\Documents\Star Swarm\Output_14_05_11_1403.txt
Version 1.10
05/11/2014 14:03
===========================================================

== Hardware Configuration =================================
GPU: AMD Radeon R9 200 Series
CPU: AuthenticAMD
AMD Phenom(tm) II X6 1090T Processor
Physical Cores: 6
Logical Cores: 6
Physical Memory: 8549400576
Allocatable Memory: 140737488224256
===========================================================


== Configuration ==========================================
API: DirectX
Scenario: ScenarioCustom.csv
User Input: Disabled
Resolution: 1920x1080
Fullscreen: True
GameCore Update: 16.6 ms
Bloom Quality: High
PointLight Quality: High
ToneCurve Quality: High
Glare Overdraw: 16
Shading Samples: 64
Shade Quality: Mid
Deferred Contexts: Disabled
Temporal AA Duration: 16
Temporal AA Time Slice: 2
Detailed Frame Info: Off
===========================================================


== Results ================================================
Test Duration: 360 Seconds
Total Frames: 20477

Average FPS: 56.88
Average Unit Count: 4661
Maximum Unit Count: 5914
Average Batches/MS: 377.14
Maximum Batches/MS: 763.65
Average Batch Count: 6663
Maximum Batch Count: 15571
===========================================================
 
Last edited:
Average Batches/MS: 377.14
Average Batch Count: 6663

I don't know what you're running but it's something that's a couple of times less intensive than what the scenario I was running.

d6cf243d_SS_cacheoff3_337.png

400 vs. 2000 average batches/MS
7000 vs. 32 000 average batches

Sugarhell also had this same issue on OCN I think, try admin permissions on the csv file, delete the old custom file etc.

You should also see the same 30-35K average batches and a fixed camera angle. Otherwise it's not working and SS is bugging out (which it admittedly does quite often)
 
I don't know what you're running but it's something that's a couple of times less intensive than what the scenario I was running.

d6cf243d_SS_cacheoff3_337.png

400 vs. 2000 average batches/MS
7000 vs. 32 000 average batches

Sugarhell also had this same issue on OCN I think, try admin permissions on the csv file, delete the old custom file etc.

You should also see the same 30-35K average batches and a fixed camera angle. Otherwise it's not working and SS is bugging out (which it admittedly does quite often)

I ran the file you gave me in the quote below.

As for your 780TI @ 1300Mhz in this quote, i'm 3 FPS off it with an R9 290 @ 1040/1350, and it can't be the same one you gave me as your FPS in that one you are in the 90's.

so which one did you give me and which one did you just run?

Replace the CustomScenario.csv in your SS folder with this: https://mega.co.nz/#!AYMGyABQ!19Mpus89txDgJ0Bmi39eevh6LMsUzs43_0lgQZSRiKY

And then run the bench with extreme settings using the custom scenario option. You should get a scenario with a fixed camera angle and fixed ship spawns (someone on here made the thing, can't remember who, but thanks to whoever it was).

It's completely different from the follow mode. With follow I've seen over 100fps on my rig.



Couldn't possibly have anything to do with CPU clocks, testing places within the game, etc.?

It's impossible to compare without knowing where and how these guys were testing.

Gamegpu.ru shows the same thing though with their testing of the latest BF4 expansion (at least it's the latest as far as I'm aware, I don't play the thing).

http://gamegpu.ru/action-/-fps-/-tps/battlefield-4-naval-strike-test-gpu.html

1080p avg/min:
780Ti: 97/74
290X M: 87/74
290X DX: 79/64

1600p avg/min:
780Ti: 60/43
290X M: 55/43
290X DX: 51/41

And they even have a youtube video of their testing routine so anyone can try it.

Of course I could always try and get BF4 working on my rig to try this stuff out myself.

However that's not really the point here, the point is that you can't universally apply a "terrible" tag on DX11 as humbug did earlier. AMD's and Nvidia's DX11 implementations are not equal at the moment. NV's is leaps and bounds above AMD's whenever you see a CPU bottleneck. Mantle might be ahead in efficiency but then again it's quite limited as far as the games library goes.
 
Last edited:
The one in my runs is the one linked in the OCN thread: http://www.overclock.net/t/1479451/...-wonder-driver-beta-release/880#post_22083320

the mega link is the same scenario as far as I can tell. No idea why yours is bugging out. (again Sugarhell on OCN had the same issue at first).

But the thing is that you should end up with very similar batch numbers as I did. 30K+ average batches, otherwise it's not the same test.

Another source is here: http://forums.overclockers.co.uk/showpost.php?p=25762826&postcount=94

Should look like this at the start of the bench:

U796bgR.png
 
The one in my runs is the one linked in the OCN thread: http://www.overclock.net/t/1479451/...-wonder-driver-beta-release/880#post_22083320

the mega link is the same scenario as far as I can tell. No idea why yours is bugging out. (again Sugarhell on OCN had the same issue at first).

But the thing is that you should end up with very similar batch numbers as I did. 30K+ average batches, otherwise it's not the same test.

Another source is here: http://forums.overclockers.co.uk/showpost.php?p=25762826&postcount=94

Should look like this at the start of the bench:

U796bgR.png

Yes that is what it looks like.

If your using the same file for both benches how is it that in the first one linked to your post directing me to it you claim to have 97 FPS, and yet in the second one your screenshot results show 59 FPS, which is pretty much what i get in Mantle, and in DirectX with the file you sent me, the reason for that is with the file you sent me the CPU is not being stressed at all in DirectX.

The only result you did actually screenshot shows your performance is no better than mine.

What do we take away from that?
 
Last edited:
What do we take away from that?

The 90s numbers you're quoting are gamegpu.ru's BF4 numbers....

And what we can take away from your scores is that your system scores similarly to mine while mine is handling 350% more batches during the whole run.

Of course we could actually make some conclusions if you were willing to do a bit of tweaking in order to get the file working correctly. Blame Oxide with this, it's not my fault that their benchmark is a buggy piece of garbage with custom scenarios. If you want a comparable result you need one with 30-35K average batches.
 
The 90s numbers you're quoting are gamegpu.ru's BF4 numbers....

And what we can take away from your scores is that your system scores similarly to mine while mine is handling 350% more batches during the whole run.

Of course we could actually make some conclusions if you were willing to do a bit of tweaking in order to get the file working correctly. Blame Oxide with this, it's not my fault that their benchmark is a buggy piece of garbage with custom scenarios. If you want a comparable result you need one with 30-35K average batches.

Right. i miss-read that then. i thought it was your SS Bench. I will hold my hands up to making a mistake there :)

What i do know is i get similar FPS to you in Mantle with a similar batch count using its Extreme preset, if we are talking about batch counts you and i are getting similar performance.

SS is designed to stress the CPU, its not a benchmark or anything like that, it is setup to push the CPU hard. really hard.

Messing about with file editing the settings simply perverts that, the most reliable way to compare anything with it is to leave it just s it is.
 
If you don't want to fix the custom scenario then here's a follow one for you to beat:

wmEG3A6.png

As I said, I've seen over 100 in the past but I can't be bothered with the inconsistent follow mode for long enough to get a really good run.

But you really should try fixing the custom scenario... Once you get it running properly (once you see 30-35K batches) it'll give really consistent results and is more stressful than follow.
 
If you don't want to fix the custom scenario then here's a follow one for you to beat:

wmEG3A6.png

As I said, I've seen over 100 in the past but I can't be bothered with the inconsistent follow mode for long enough to get a really good run.

But you really should try fixing the custom scenario... Once you get it running properly (once you see 30-35K batches) it'll give really consistent results and is more stressful than follow.

I can't get near that.

But why is your Average Batch Count half in this one than what it was in the one quoted below, the one where you only get 59 FPS?

@ 59 FPS your running 31K Batch count, @ 96 FPS your only running 15K

i'm running 57 FPS with an Avr Batch Count of 24K, i ran it 3 times and could not get it to go over 57 FPS, or get the Batch count much more under 24K.

===========================================================
Oxide Games
Star Swarm Stress Test - ©2013
C:\Users\#\Documents\Star Swarm\Output_14_04_11_1444.txt
Version 1.10
04/11/2014 14:44
===========================================================

== Hardware Configuration =================================
GPU: AMD Radeon R9 200 Series
CPU: AuthenticAMD
AMD Phenom(tm) II X6 1090T Processor
Physical Cores: 6
Logical Cores: 6
Physical Memory: 8549400576
Allocatable Memory: 140737488224256
===========================================================


== Configuration ==========================================
API: Mantle
Scenario: ScenarioFollow.csv
User Input: Disabled
Resolution: 1920x1080
Fullscreen: True
GameCore Update: 16.6 ms
Bloom Quality: High
PointLight Quality: High
ToneCurve Quality: High
Glare Overdraw: 16
Shading Samples: 64
Shade Quality: Mid
Deferred Contexts: Disabled
Temporal AA Duration: 16
Temporal AA Time Slice: 2
Detailed Frame Info: C:\Users\#\Documents\Star Swarm\FrameDump_14_04_11_1444.csv
===========================================================


== Results ================================================
Test Duration: 360 Seconds
Total Frames: 20453

Average FPS: 56.81
Average Unit Count: 4635
Maximum Unit Count: 5596
Average Batches/MS: 971.75
Maximum Batches/MS: 3639.21
Average Batch Count: 24102
Maximum Batch Count: 146378
===========================================================

I don't know what you're running but it's something that's a couple of times less intensive than what the scenario I was running.

d6cf243d_SS_cacheoff3_337.png

400 vs. 2000 average batches/MS
7000 vs. 32 000 average batches

Sugarhell also had this same issue on OCN I think, try admin permissions on the csv file, delete the old custom file etc.

You should also see the same 30-35K average batches and a fixed camera angle. Otherwise it's not working and SS is bugging out (which it admittedly does quite often)
 
I didn't say they didn't, I'm just interested in how much the difference is.
How much profit do they make on the £400 hardware? How much does it cost to build, test, package and ship? Also, not every console gamer will have a subscription.
If there wasn't money in PCs I doubt Microsoft would've gotten into the market at all. If they don't get anything from it why have made DirectX at all and why would they have kept it up?




Exactly everyone makes it sound like consoles and PCs are mutually exclusive. I know a number of people with both (maybe not all Microsoft consoles, but again DX12 being good on the PC won't change that).
It's like people think Microsoft don't know how to make money...

They don't make much money from the hardware and usually make a loss at launch. The profits are made from games, online service's and peripheral's.
 
Last edited:
Back
Top Bottom