Ars Technica took it upon themselves to benchmark Metal in the latest OSX El Capitan release. Even though OpenGL on Mac OSX is not considered to be on par with its Linux counterparts, which is probably due to the driver situation until recently, it pulls ahead of Metal in many situations.
Image Credit: Ars Technica
Unlike the other graphics APIs, Metal uses the traditional binding model. Basically, you have a GPU object that you attach your data to, then call one of a handful of “draw” functions to signal the driver. DirectX 12, Vulkan, and Mantle, on the other hand, treat work like commands on queues. The latter model works better in multi-core environments, and it aligns with GPU compute APIs, but the former is easier to port OpenGL and DirectX 11 applications to.
Ars Technica notes that faster GPUs, such as the NVIDIA GeForce GTX 680MX, show higher gains than slower ones. Their “best explanation” is that “faster GPUs can offload more work from the CPU”. That is pretty much true, yes. The new APIs are designed to keep GPUs loaded and working as much as possible, because they really do sit around doing nothing a lot. If you are able to keep a GPU loaded, because it can't accept much load in the first place, then there is little benefit to decreasing CPU load or spreading out across multiple cores.
Granted, there are many ways that benchmarks like these could be incorrectly used. I'll assume that Ars Technica and GFXBench are not making any simple mistakes, though, but it's good to be critical just in case.
Who names a lightweight API
Who names a lightweight API Vulkan?
Yeah. At least the Glide API
Yeah. At least the Glide API was appropriately named.
Haven’t you heard of Mantle?
Haven’t you heard of Mantle? Vulkan, the German word for volcano actually makes sense as a name for a successor that builds upon Mantle.
Besides, it is also simply called glNext.
Let’s be honest – they’ve
Let’s be honest – they’ve named it Vulkan simply because Leonard Nimoy passed away shortly before Khronos group managed to introduce new Mantle-based OpenGL to worldwide audience. I bet your ass that if Leonard Nimoy was still alive (at least by then), new OGL would’ve been called something else completely different altogether.
No before Roddenberry
No before Roddenberry borrowed the name Vulcan the Roman god of fire’s name for his fictitious space people/civilization, the name was not associated with a science fiction series. It’s more of a geological reference to Mantle(the Geological term), because Vulkan(German spelling) is what is created by the Mantle when it reaches the surface. Khronos had the name Vulkan trademarked before Nimoy’s death.
From Wikipedia:
“The Khronos Group began a project to create a next generation graphics API in summer 2014. At SIGGRAPH 2014 the project was publicly announced with a call for participants.[6]
According to the US Patent and Trademark Office, the trademark for Vulkan was filed on February 19, 2015.[17]
Vulkan was formally named and announced at Game Developers Conference 2015, although speculation and rumors centered around a new API existed beforehand and referred to it as ‘glNext’.[18]”
[6,17,18] see Vulkan(API) Wikipedia entry.
1. It’s spelled “Vulkan”, not
1. It’s spelled “Vulkan”, not “Vulcan”.
2. “Vulkan” with “k” is the name of the planet and the name of the race.
3. They called it “Vaulkan” with “k” and nothing other, and they did almost right after Leonard Nimoy has passed away.
4. Everything indicates that they’ve gone with that naming scheme simply to commemorate Leonard Nimoy’s, “Spock’s”, legacy. Nothing less or more than just that.
Your argument is invalid.
Nope. They filed for that
Nope. They filed for that name prior to Mr. Nimoy’s death. Just a very unfortunate coincidence.
It’s Vulcan in the Roman God
It’s Vulcan in the Roman God of Fire, and Vulkan in German, and The place and date where the API name was formally announced has nothing to do with the date the name Vulkan(API) was Trademarked(February 19, 2015), the term was Trademarked before Nimoy’s death, and considering the legal work involved the Name Vulkan was decided on by Khronos long before the copyright paperwork was filed!
Nimoy died on February 27, 2015, the trademark was filed before his death! It’s spelled Vulcan in the Star Trek TV series also, oh how those TV writers like to borrow from Gerco/Roman/Norse mythology! You have been Mind melding with vegetables a little too much!
Because the lightweight(not
Because the lightweight(not really) Vulkan graphics API is made from mostly AMD’s Mantle, so in keeping with the Geological reference from AMD’s use of the name Mantle for its close to the metal(hardware) Vulkan graphics API. Apple is going to have to rework its Metal(API) to make it work more like Vulkan, to be able to take more advantage of asynchronous compute, hopefully Apple will adopt Vulkan also, in addition to its Metal.
Vulkan is more of a HSA style API that is able to utilize SPIR-V IL to accelerate more than just Grophical workloads on GPUs. OpenCL has always used SPIR/SPIR-V(latest version). (A)
(A) from wikipedia entry on SPIR.
“Standard Portable Intermediate Representation (SPIR) is an intermediate language for parallel compute and graphics by Khronos Group, originally developed for use with OpenCL. The current version, SPIR-V, was announced in March 2015.”
Need AMD performance numbers
Need AMD performance numbers on more recent iMacs.
So I assume this is about
So I assume this is about OpenGL 4.1? Just wondering how much OpenGL 4.5 + AZDO would do the difference.
I’ve followed Ars technica
I’ve followed Ars technica for ~5 years now. They’re niche has definitely been moving away from the more technical PC hardware, micro-architecture, and benchmarking type of content. I’m not saying benchmarks they publish are wrong, but I’d look for verification from other sites.
Ars Technica was never on the
Ars Technica was never on the same level as the Microprocessor Report, or other professional trade journals, and no publications outside of this and other pay-walled professional publications are able to provide the proper professional analysis of CPU/other processor architectures. Benchmarking is far from an exact science, the way benchmarking is reported and Used/Gamed by some online “Technology” sites.
Only the proper scientific peer reviewed benchmarking can be trusted where they make sure the results are exact and reproducible by other indipendent testers. Benchmarking software needs to be properly compiled on the testing platform, using the same testing hardware, and the exact same compiler flags and optimizations. It’s best to spend the money and subscribe to the professional publications.