Why AMD’s new focus on ARM means a pivot away from the original focus of HSA.
On Monday AMD announced plans for one of the more interesting and radical shifts in architecture since the announcement of the APU: bringing ARM and x86 together on the same pin compatible chip, as well as the eventual release of an AMD-designed ARM.
The integration of ARM into AMD under the banner of Project Skybridge — which sounds suspiciously like Intel’s Skylake — and what it means for the future of HSA was a question raised in an earlier piece published on VR-Zone, and solicited a fierce discussion in the article’s comment section. This piece seeks to clarify some points and expand on why AMD is pivoting towards ARM and away from HSA, while still integrating HSA onto its new silicon, with Skylake.
First, its important to differentiate between HSA on the desktop and HSA on mobile. In the mobile world, HSA and heterogeneous compute seem to have moved much further along than in the desktop space. MediaTek (one of the HSA Foundation’s founding members) *is making great progress and has some clear wins in the space with things like CorePilot, its Heterogeneous Multi-Processing (HMP) control software for big.LITTLE. In contrast, AMD has not had the same sort of success in the desktop world. While AMD has long extolled the virtues of HSA, its much-awaited Kaveri APU was released earlier this year but it didn’t live up to its full potential because HSA-enabled apps are non-existent. The momentum simply isn’t there.
On the desktop will the momentum ever be there to fully implement HSA? From a cost-benefit analysis maybe, but only for certain applications. AMD holds a fraction of the market share that Intel commands, and its CPUs (Firepro GPUs are another story) aren’t used in any meaningful numbers in the PC productivity space. For a developer it just isn’t worth it to port in HSA to productivity apps, though gaming is a different story.
So AMD has come to the realization that HSA in its current form is not working. The original advertised focus on the language, ambidextrous computing via the simultaneous addressing of both CPU and GPU cores on the chip, is being refocused now to include ARM. Though HSA isn’t going anywhere fast, the HSA we know has been pushed from the limelight in favour of ambidextrous x86-ARM computing.
The interest in ARM is rising, and vendors are experimenting with putting it on the PC; AMD is trying to do it in a big way. But as discussed before, AMD just simply doesn’t have the market share to dictate radical changes like this (maybe that contributes to the near-junk status of its credit rating). It’s as if the company figures that its one coup of getting Intel to also place GPU cores on the CPU silicon can be repeated.
AMD’s next move is to expand upon what it’s shared so far. A likely venue for this will be Computex next month in Taipei.
First, its important to differentiate between HSA on the desktop and HSA on mobile. In the mobile world, HSA and heterogeneous compute seem to have moved much further along than in the desktop space. MediaTek (one of the HSA Foundation’s founding members) *is making great progress and has some clear wins in the space with things like CorePilot, its Heterogeneous Multi-Processing (HMP) control software for big.LITTLE. In contrast, AMD has not had the same sort of success in the desktop world. While AMD has long extolled the virtues of HSA, its much-awaited Kaveri APU was released earlier this year but it didn’t live up to its full potential because HSA-enabled apps are non-existent. The momentum simply isn’t there.
On the desktop will the momentum ever be there to fully implement HSA? From a cost-benefit analysis maybe, but only for certain applications. AMD holds a fraction of the market share that Intel commands, and its CPUs (Firepro GPUs are another story) aren’t used in any meaningful numbers in the PC productivity space. For a developer it just isn’t worth it to port in HSA to productivity apps, though gaming is a different story.
So AMD has come to the realization that HSA in its current form is not working. The original advertised focus on the language, ambidextrous computing via the simultaneous addressing of both CPU and GPU cores on the chip, is being refocused now to include ARM. Though HSA isn’t going anywhere fast, the HSA we know has been pushed from the limelight in favour of ambidextrous x86-ARM computing.
The interest in ARM is rising, and vendors are experimenting with putting it on the PC; AMD is trying to do it in a big way. But as discussed before, AMD just simply doesn’t have the market share to dictate radical changes like this (maybe that contributes to the near-junk status of its credit rating). It’s as if the company figures that its one coup of getting Intel to also place GPU cores on the CPU silicon can be repeated.
AMD’s next move is to expand upon what it’s shared so far. A likely venue for this will be Computex next month in Taipei.
Read More: http://ift.tt/1mB2fDN
via VRForums | Singapore Technology Lifestyle Forums - News around the web! http://ift.tt/1mAKhWe
No comments:
Post a Comment