Jump to content

AMD Mantle Investigation

LinusTech

Meh unless something like mantle is a standard backed up by the competition too this is gonna be a slow mess.In theory is great but trouble on implementation as usual.

I honestly wish microsoft would come up with a standard directx12 low level api in the future if the benefit is really there.And i think it is at least on the cpu side so why not have this great idea as standard goes beyond my understanding..

Soon there will be SteamOS with OGL and even more confusion they just release mantle for windows and now what ?release another buggy one for linux?too messy if you ask me.Plus the problem with bottlenecks of the cpu is because dev's dont bother optimizing the code on most of the PC games anymore except a few like DICE ie: BF3 runs on my i5 4570 at 100% while on BF4 which is more graphic intensive the cpu usage is 50-80% depending on map(ultra ofc) so it can be done if they care they dont need any mantle.

When i saw StarSwarm i tought this would be great for Rome II Total War since it runs like sh4t but then i realized CA has no clue about game engines and optimizations,hell everytime i run that game theres always a core maxing out 100% like great multicore optimizations there guys(CA).The API is the least of the problems IMO.

Link to comment
Share on other sites

Link to post
Share on other sites

Just watched your video and decided to run Star Swarm Stress Test to compare the performance. Mantle is running slower that D3D is. It's a bit dissapointig to be honest. That also concludes my consideration for my next graphics card.

Link to comment
Share on other sites

Link to post
Share on other sites

What card did you run it on?

System CPU : Ryzen 9 5950 doing whatever PBO lets it. Motherboard : Asus B550 Wifi II RAM 80GB 3600 CL 18 2x 32GB 2x 8GB GPUs Vega 56 & Tesla M40 Corsair 4000D Storage: many and varied small (512GB-1TB) SSD + 5TB WD Green PSU 1000W EVGA GOLD

 

You can trust me, I'm from the Internet.

 

Link to comment
Share on other sites

Link to post
Share on other sites

I ran it on 7770 with a little overclock. My CPU is i7 2600k 4.3GHz.

Link to comment
Share on other sites

Link to post
Share on other sites

Curious. Well it's still beta unless you're buying soon don't give up on it already until it hits non beta release

System CPU : Ryzen 9 5950 doing whatever PBO lets it. Motherboard : Asus B550 Wifi II RAM 80GB 3600 CL 18 2x 32GB 2x 8GB GPUs Vega 56 & Tesla M40 Corsair 4000D Storage: many and varied small (512GB-1TB) SSD + 5TB WD Green PSU 1000W EVGA GOLD

 

You can trust me, I'm from the Internet.

 

Link to comment
Share on other sites

Link to post
Share on other sites

Yes! Better Star Citizen performance :D

| Case: NZXT Tempest 210 | CPU: Intel Core i5 3570K @ 3.9 Ghz | GPU: ASUS ROG STRIX GTX 1070 | RAM: Crucial Ballistix Tactical 8GB |

| Mouse: Zowie FK1 | Monitor: Acer 21.5' | Keyboard: CoolerMaster Stealth w/ Brown Switches |

#KilledMyWife - #LinusButtPlug - #1080penis

 

Link to comment
Share on other sites

Link to post
Share on other sites

Yes! Better Star Citizen performance :D

. This is a man who can find the good in most any article
System CPU : Ryzen 9 5950 doing whatever PBO lets it. Motherboard : Asus B550 Wifi II RAM 80GB 3600 CL 18 2x 32GB 2x 8GB GPUs Vega 56 & Tesla M40 Corsair 4000D Storage: many and varied small (512GB-1TB) SSD + 5TB WD Green PSU 1000W EVGA GOLD

 

You can trust me, I'm from the Internet.

 

Link to comment
Share on other sites

Link to post
Share on other sites

Seeing as I have a 7790 this could really help me benefit by using mantle. I hear Thief is supporting it. I'm eager to see how things will turn out when I run mantle supported games.

Link to comment
Share on other sites

Link to post
Share on other sites

Hard to say to be honest. G-SYNC requires a fancy new monitor that isn't available yet.

 

Mantle requires a fancy new graphics card or APU (to really get the most of it), a shoddy CPU, and games that aren't available yet.

You don't need a shoddy CPU as it turns out, even the extreme edition six core Intel CPUs showed a measurable performance improvement in Anandtech's benchmarks and yours in fact.

You also need to keep in mind that mobile platforms like gaming notebooks commonly do actually have ~2.5Ghz processors paired with GTX 680, GTX 660 & 7870 equivalents.

So more often than not they're CPU bound.

And it's only going to get worse with each manufacturing node GPU performance doubles per mm², per watt & per dollar while CPU performance only increases by a few percentages, especially single thread performance.

Link to comment
Share on other sites

Link to post
Share on other sites

Oh I thought those were going to be the ones that don't make sense.

Thanks for posting though.

No? As stated the ones that didn't make sense were all within BF4

Link to comment
Share on other sites

Link to post
Share on other sites

No? As stated the ones that didn't make sense were all within BF4

what system was used for the RTS benchmarks? It doesn't say in the link you provided :-)

Motherboard - Gigabyte P67A-UD5 Processor - Intel Core i7-2600K RAM - G.Skill Ripjaws @1600 8GB Graphics Cards  - MSI and EVGA GeForce GTX 580 SLI PSU - Cooler Master Silent Pro 1,000w SSD - OCZ Vertex 3 120GB x2 HDD - WD Caviar Black 1TB Case - Corsair Obsidian 600D Audio - Asus Xonar DG


   Hail Sithis!

Link to comment
Share on other sites

Link to post
Share on other sites

3960X stock 290X stock

Danke

Motherboard - Gigabyte P67A-UD5 Processor - Intel Core i7-2600K RAM - G.Skill Ripjaws @1600 8GB Graphics Cards  - MSI and EVGA GeForce GTX 580 SLI PSU - Cooler Master Silent Pro 1,000w SSD - OCZ Vertex 3 120GB x2 HDD - WD Caviar Black 1TB Case - Corsair Obsidian 600D Audio - Asus Xonar DG


   Hail Sithis!

Link to comment
Share on other sites

Link to post
Share on other sites

3960X stock 290X stock

did you noticed much of a difference in going from a 6 core to a quad core?

Motherboard - Gigabyte P67A-UD5 Processor - Intel Core i7-2600K RAM - G.Skill Ripjaws @1600 8GB Graphics Cards  - MSI and EVGA GeForce GTX 580 SLI PSU - Cooler Master Silent Pro 1,000w SSD - OCZ Vertex 3 120GB x2 HDD - WD Caviar Black 1TB Case - Corsair Obsidian 600D Audio - Asus Xonar DG


   Hail Sithis!

Link to comment
Share on other sites

Link to post
Share on other sites

did you noticed much of a difference in going from a 6 core to a quad core?

Remember in the video? Results were very weird and I don't wish to report them at this time, I await a new driver from AMD

Link to comment
Share on other sites

Link to post
Share on other sites

Remember in the video? Results were very weird and I don't wish to report them at this time, I await a new driver from AMD

''No? As stated the ones that didn't make sense were all within BF4''

 

I was asking about the RTS game. You stated that the bad results were from BF4. at least that's what I understood :P

Motherboard - Gigabyte P67A-UD5 Processor - Intel Core i7-2600K RAM - G.Skill Ripjaws @1600 8GB Graphics Cards  - MSI and EVGA GeForce GTX 580 SLI PSU - Cooler Master Silent Pro 1,000w SSD - OCZ Vertex 3 120GB x2 HDD - WD Caviar Black 1TB Case - Corsair Obsidian 600D Audio - Asus Xonar DG


   Hail Sithis!

Link to comment
Share on other sites

Link to post
Share on other sites

but... i only tested the 1 processor with star swarm lol

 

I think the point was made with the tests I ran and I don't feel it would have added much to put more platforms in the mix - honestly I didn't find processor to make as much of an impact as most people(including myself) expected

Link to comment
Share on other sites

Link to post
Share on other sites

but... i only tested the 1 processor with star swarm lol

 

I think the point was made with the tests I ran and I don't feel it would have added much to put more platforms in the mix - honestly I didn't find processor to make as much of an impact as most people(including myself) expected

That is really strange....considering that Mantle's thing is alleviating the CPU bottleneck....

What an odd product launch

Motherboard - Gigabyte P67A-UD5 Processor - Intel Core i7-2600K RAM - G.Skill Ripjaws @1600 8GB Graphics Cards  - MSI and EVGA GeForce GTX 580 SLI PSU - Cooler Master Silent Pro 1,000w SSD - OCZ Vertex 3 120GB x2 HDD - WD Caviar Black 1TB Case - Corsair Obsidian 600D Audio - Asus Xonar DG


   Hail Sithis!

Link to comment
Share on other sites

Link to post
Share on other sites

That is really strange....considering that Mantle's thing is alleviating the CPU bottleneck....

What an odd product launch

As far as I can tell its more based on how the CPU sits in the overall equation. Sure you will see improvements in % of performance gained when you're on a slower CPU but its not THAT huge... and no one is realistically running a 290X with a 2Ghz CPU

 

its more the fact that mantle is taking global steps to alleviate CPU bottlenecks, global steps that help all different proc regardless of Ghz - albeit some more than others

Link to comment
Share on other sites

Link to post
Share on other sites

As far as I can tell its more based on how the CPU sits in the overall equation. Sure you will see improvements in % of performance gained when you're on a slower CPU but its not THAT huge... and no one is realistically running a 290X with a 2Ghz CPU

 

its more the fact that mantle is taking global steps to alleviate CPU bottlenecks, global steps that help all different proc regardless of Ghz - albeit some more than others

Did you notice a difference in graphical fidelity with Mantle? I know some people did. Some where saying that the difference in FPS is caused by that inconsistency compared to D3D

Motherboard - Gigabyte P67A-UD5 Processor - Intel Core i7-2600K RAM - G.Skill Ripjaws @1600 8GB Graphics Cards  - MSI and EVGA GeForce GTX 580 SLI PSU - Cooler Master Silent Pro 1,000w SSD - OCZ Vertex 3 120GB x2 HDD - WD Caviar Black 1TB Case - Corsair Obsidian 600D Audio - Asus Xonar DG


   Hail Sithis!

Link to comment
Share on other sites

Link to post
Share on other sites

Did you notice a difference in graphical fidelity with Mantle? I know some people did. Some where saying that the difference in FPS is caused by that inconsistency compared to D3D

No but I didn't look that closely as again this is an issue and will be fixed in a patch at some point.

 

Everyone seems to easily forget that it is still in beta, put the pitchforks back on the rack until it is release... at which point if there are still these problems by all means grab your pitchforks.

 

I also doubt it would have created that much of a performance gap - but i guess that is yet to be seen once they fix it

Link to comment
Share on other sites

Link to post
Share on other sites

No but I didn't look that closely as again this is an issue and will be fixed in a patch at some point.

 

Everyone seems to easily forget that it is still in beta, put the pitchforks back on the rack until it is release... at which point if there are still these problems by all means grab your pitchforks.

 

I also doubt it would have created that much of a performance gap - but i guess that is yet to be seen once they fix it

I am understanding that it is in BETA, I wasn't holding it against them but I was wondering if the discrepancy in fidelity was caused by certain HW combinations resulting in a bug or not.

 

Any word when the next drivers are supposed to be out? (probably an NDA thing) Hopefully they are legit drivers and not BETA ones. :D

Motherboard - Gigabyte P67A-UD5 Processor - Intel Core i7-2600K RAM - G.Skill Ripjaws @1600 8GB Graphics Cards  - MSI and EVGA GeForce GTX 580 SLI PSU - Cooler Master Silent Pro 1,000w SSD - OCZ Vertex 3 120GB x2 HDD - WD Caviar Black 1TB Case - Corsair Obsidian 600D Audio - Asus Xonar DG


   Hail Sithis!

Link to comment
Share on other sites

Link to post
Share on other sites

Frankly I don't care about Mantle in particular but I DO care about the draw-call issue being fixed... and if Mantle is the vehicle that carries us there then I'm on board!

 

If Nvidia or someone else comes up with a different solution well I'll probably be on-board with that as well, especially if it doesn't push segmentation and increased developer workload

Link to comment
Share on other sites

Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now

×