Mantle Feedback/Bugs

[Rusty0611]

You are not here to help us try to fix bugs or provide feedback. This thread is not for posting downfalls of mantle or caveats of mantle or discussing benchmark results in context of those assumed downfalls.

I posted a freezing/white screens issue I had, and it got overlooked and got lost between these unnecessary (for this thread not in general) posts, so I am angry, not being able to get help from your fellow peers if they've already resolved the issue does that to you. So you are more than welcome to move on to Mantle Benchmark Thread and post your points and remarks there, I haven't read them completely since they were out of context of this thread, but some of your points have face value and may be worth discussing really but not just here.

Matt posted the benchmark here not me buddy. Therefore the discussion of it was here.

Find fault / scrutinise / not representative. All the same. Its not my benchmark, but im just explaining the context of it and why its not reliable to compare it to something else.

If you have no opinion having not used it then >> time to leave the thread. Perhaps the Mantle Benchmark thread would be better for you. If you actually check that out, you'll see real world examples of the performance difference in exactly the scenario's you mentioned. I saw a 14% increase with an i7 and a single 290 card in the Mantle benchmark thread.

You posted a benchmark here and it was commented upon here. If you didn't want that to happen here then you should have posted it in the other thread. Simples .

My apologies if my posts were not under the umbrella of 'feedback'.

 

[uksoldierboy]

It hasn't really though. It was all about Mantle and there isn't a Mantle GD thread. This one fits the bill.

Feedback is a wide ranging term so to me encompasses everything. The mods directed Mantle discussion here after the old thread closed.

As Andy said: "it takes two to tango" .

Here's my "user" data on Mantle so far, if it takes two to tango and all..your go.

DX

Spoiler

== Configuration ==========================================
API: DirectX
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: Off
================================================== =========


== Results ================================================
Test Duration: 360 Seconds
Total Frames: 15692

Average FPS: 43.58
Average Unit Count: 4097
Maximum Unit Count: 5336
Average Batches/MS: 749.27
Maximum Batches/MS: 1862.05
Average Batch Count: 18836
Maximum Batch Count: 121944

Mantle

Spoiler

== 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: Off
================================================== =========


== Results ================================================
Test Duration: 360 Seconds
Total Frames: 26063

Average FPS: 72.39
Average Unit Count: 4658
Maximum Unit Count: 6024
Average Batches/MS: 1120.25
Maximum Batches/MS: 3426.47
Average Batch Count: 17820
Maximum Batch Count: 130588

A 66.11% increase for me on the "follow" bench.

Apart from my star swarm comparison of course

Sorry Atomic lol

 

[Rusty0611]

Here's my "user" data on Mantle so far, if it takes two to tango and all..your go.

DX

Spoiler

== Configuration ==========================================
API: DirectX
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: Off
================================================== =========


== Results ================================================
Test Duration: 360 Seconds
Total Frames: 15692

Average FPS: 43.58
Average Unit Count: 4097
Maximum Unit Count: 5336
Average Batches/MS: 749.27
Maximum Batches/MS: 1862.05
Average Batch Count: 18836
Maximum Batch Count: 121944

Mantle

Spoiler

== 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: Off
================================================== =========


== Results ================================================
Test Duration: 360 Seconds
Total Frames: 26063

Average FPS: 72.39
Average Unit Count: 4658
Maximum Unit Count: 6024
Average Batches/MS: 1120.25
Maximum Batches/MS: 3426.47
Average Batch Count: 17820
Maximum Batch Count: 130588

A 66.11% increase for me on the "follow" bench.

Sorry Atomic lol

Eh hmmm... 'wrong thread' .

I'm particularly interested in BF4. Depending on the success of Mantle I may consider it when I next upgrade my GPU. At the moment I just buy the best for me at that time. Mantle doesn't factor into this decision at the moment but it could well do in the future.

 

[shankly1985]

Am not sure why they is so much negativity towards Mantle?

 

[Rusty0611]

Am not sure why they is so much negativity towards Mantle?

Where on earth are you seeing negativity?

 

[LtMatt]

You posted a benchmark here and it was commented upon here. If you didn't want that to happen here then you should have posted it in the other thread. Simples .

My apologies if my posts were not under the umbrella of 'feedback'.

No problem with talking about the benchmark, it was posted here to be discussed. Its also something us Mantle users can try and copy and replicate and see what results we get. However you spoke about wanting to see real world results, i just informed you that you've been looking in the wrong place for them and derailing the thread somewhat in the process.

 

[Rusty0611]

No problem with talking about the benchmark, it was posted here to be discussed. Its also something us Mantle users can try and copy and replicate and see what results we get. However you spoke about wanting to see real world results, i just informed you that you've been looking in the wrong place for them and derailing the thread somewhat in the process.

No that wasn't what I said. I said that as an addendum to my larger more general point about graph dropping and representation.

 

[DamnedLife]

It wasn't negativity shankly, it was a discussion of validity of benchmarking mantle and two benchmark being comparable or not, but then it took a nasty turn. This back and forth is unnecessary, doesn't matter who started it.

Also since mantle is already out there and appreciated by some, there is no point in sharing a benchmark "somehow" showing it is unnecessary, that somehow part doesn't matter as the argument itself is irrelevant.

 

[ICDP]

I have no opinion having not used it but looking at the data it's difficult to see how much benefit it's going to have for your average high end GPU user with an overclocked i5/i7 who uses 1080/1440 on max settings.

Do you believe honestly believe a 10%-40% performance (depending on system) boost is no benefit? HardOCP showed that even at 2560x1600 there are tangible benefits to mantle with an overclocked high i7 3770k at 4.8GHz. Their testing shows a 20%+ improvement at 1920x1080 in BF4 MP.

My own tests show up 100%-167% improvement in Star Swarm depending on the scenario. This is with an i7 4770k at 4.4GHz and R9 290X at 1100/1300.

Mantle is an alpha/beta and as a proof of concept it has more than shown it's worth. Clearly there is far more to come. There are far too many people and review sites that show very tangible 10%-40% + improvements in BF4 even on high end hardware.

 

[DamnedLife]

BTW back to topic... After 3 clean installs later, I can say that messing with some BF4 settings (VSync, MSAA and terrain detail - haven't singled out which) somehow may end up breaking the driver, forcing you to clean install.

 

[Rusty0611]

Do you believe honestly believe a 10%-40% performance (depending on system) boost is no benefit?

Anything is of benefit. What I meant, is that a benefit weighty enough to throw it into my own purchase decisions?

10% is neither here nor there IMO but anything up towards 20% becomes in my eyes more worthwhile.

 

[Diagro]

Rusty. The gains are in excess of 30% In my system. Plus its like night and day vs DX.....

You haven't used it, and are ignoring the data.

 

[Woodsta888]

Anything is of benefit. What I meant, is that a benefit weighty enough to throw it into my own purchase decisions?

10% is neither here nor there IMO but anything up towards 20% becomes in my eyes more worthwhile.

Well, why don't you just wait for some dedicated website reviews to tell you what you need to know instead of tearing apart some forum members attempt to produce some benchmark that we can, at this stage understand at beta?

Is this thread set up to inform you of at least a 20% uplift in performance before YOU. And why at the beta stage would you expect that? ** reign it in a bit please ** . You will find you answer out as AMD develops. But you seem to expect to make YOUR purchase from beta releases.

Bet you bougfht an i7 at the BF4 beta release.

 

[Suarez7]

It also depends on if you are going to play the titles that use it. I wouldn't base my purchase decision on it up to now but in the future that could change if more and more games start using it and nvidia don't have a counter.

I'm impressed with some of the benchmarks so far in this thread and it was obvious there would be a few niggles in the 1st Mantle driver, hopefully a smooth and fast process to get them all ironed out

 

[Rusty0611]

Rusty. The gains are in excess of 30% In my system. Plus its like night and day vs DX.....

You haven't used it, and are ignoring the data.

You've got crossfire. The gains will be bigger due to more CPU / DX limiting. I was looking at single GPU.

So no, I'm not ignoring the data you just didn't read what I said properly .

 

[shankly1985]

Anything is of benefit. What I meant, is that a benefit weighty enough to throw it into my own purchase decisions?

10% is neither here nor there IMO but anything up towards 20% becomes in my eyes more worthwhile.

You keep saying 10% ?? Am seeing much better than that, my frame rate on Ultra No MSAA went from 85fps to 120locked on Shanghai map.
This was standing in the same spot DirectX vs Mantle posted the images in the last thread.

Want to see them again?

Edit
I see you taking Single GPU, I will go and switch of Crossfire and see what I gain.

 

[Rusty0611]

Well, why don't you just wait for some dedicated website reviews to tell you what you need to know instead of tearing apart some forum members attempt to produce some benchmark that we can, at this stage understand at beta?

Is this thread set up to inform you of at least a 20% uplift in performance before YOU. And why at the beta stage would you expect that? ****. You will find you answer out as AMD develops. But you seem to expect to make YOUR purchase from beta releases.

Bet you bougfht an i7 at the BF4 beta release.

** Rusty, you should know better than to respond like that **


I'm not tearing apart anyone. I never said or implied that the thread was set up to inform me of anything - I'm just saying this is the point I would become more interested. I'm not expecting anything from it - where did you get that from?

I don't envy AMD, thanks - I had CF 7950s before my 780 so you also made the mistake of assuming somebody that has a nVidia GPU is a brand fanboy. I'm the opposite - I despise brand defending / attacking. I find it completely pathetic.

p.s. I've had an i7 since about 10 months after BF3 came out

You keep saying 10% ?? Am seeing much better than that, my frame rate on Ultra No MSAA went from 85fps to 120locked on Shanghai map.
This was standing in the same spot DirectX vs Mantle posted the images in the last thread.

Want to see them again?

Edit
I see you taking Single GPU, I will go and switch of Crossfire and see what I gain.

This is agonising. I didn't say the gains are 10% I said that if they are 10% then that would be neither here nor there for me personally. 20% becomes more interesting.

I really struggle in this sub forum sometimes.

 

[maxilive]

Anything is of benefit. What I meant, is that a benefit weighty enough to throw it into my own purchase decisions?

10% is neither here nor there IMO but anything up towards 20% becomes in my eyes more worthwhile.

Perfect comment!

For all the bickers that goes of for Nvidia vs AMD and these dumb features. I think Mantle is pretty damn good for a free boost for modern cards and it does present some interesting possibilities for the future, if they are used or not that is another question that needs to be answered. But does this instantly make AMD the better card? No, it just gives it that one extra pro when it comes to comparing a card. You should not be buying a card for Mantle though as it is a good year or so off before it becomes an actual benefit and not just a pro point when it comes to purchasing.

Its the same with Nvidia shadow play, for me that seems like a better feature as I would make more use out of streaming/video recording than I would have Mantle atm (I don't play much BF4). But it isn't a pro for me when it comes to buying a card as OBS and such do a better job for my needs.

Mantle should not sell cards and anyone who comments out of jealousy or buyers remorse of their card is just dumb. Every card has it pros and cons so just go and post about the pros and leave all the negative and comparison junk alone.

This is from someone who owns a 290 who thinks it is a decent card and Mantle does sound interesting in the future. I'm also not enjoying the reference cooler noise though which seems to be that tiny bit higher again with the 14.1 drivers compared to the previous 13 ones.

 

[DamnedLife]

Anything is of benefit. What I meant, is that a benefit weighty enough to throw it into my own purchase decisions?

10% is neither here nor there IMO but anything up towards 20% becomes in my eyes more worthwhile.

Now I am not going to argue how 10% is gathered, and assume it is correct for your setup and resolution, cpu etc. It is "a number" for JUST ONE GAME, and not going to discuss that game being designed Dx in mind or not and assume it is perfectly Mantle capable and optimized to fullest, and I am going to repeat this again, that perfectly correct number is JUST FOR ONE GAME ONLY.

Now I say that StarSwarm results are marginally higher knowing full well it is not a game alpha or anything, but its just a tech demo.

Now that makes two things running on Mantle giving huge difference in percentile gains. Third is on the way, Thief, and STILL it shouldn't be the deciding factor either.

For SS; going into an alpha stage for an actual game, it is unknown what will they do about marginal differences;
1- They may do two version like game where Dx has classic motion blur, Mantle has cinematic motion blur etc.
2- Nerf the tech demo stuff and actually release a game, same settings and effects across Dx and Mantle, which btw they strongly go against it
3- Something I haven't come up with but they have.

All speculation for the actual game atm, but still you can't deny the huge gains, hugely different to 10% you mention.

Considering Mantle is still in early phase, BF4 may have just done CPU optimizations (SS too and nothing more added on top) but they may slowly make progress in GPU optimizations and make even much higher gains for games. And consider that increase varies even in the same game environment for different specs, and my "10% is perfectly true representative" is JUST an assumption, because it isn't 10% for everyone you can't say Mantle gain for BF4 is 10%

 

[andybird123]

Yes if someone wants to scrutinise something or find fault with it, the responsibility lies with them to do so by reading the linked article.



Interested to hear both Andy's and Rustys feedback on Mantle.

I'm not suffering through installing bf4 just to see, particularly as I have one of the older cards that isnt properly supported, ive totally given up on bf4 so thief will probably be my first opportunity to evaluate assuming it allows dual install

That also doesnt preclude me in having an opinion on whether mantle will have an effect on future purchase decisions or recomendations, or in how mantle stacks up to the marketing claims