Vik


CPU Design Guru Jim Keller Joins Intel; Completes CPU Grand Tour

CPU Design Guru Jim Keller Joins Intel; Completes CPU Grand Tour

For long-time AnandTech readers, Jim Keller is a name many are familiar with. The prolific microarchitectural engineer has been involved in a number of high-profile CPU & SoC projects over the years, including AMD’s K8 and Zen CPUs and Apple’s early A-series SoCs. Now after a stint over as Tesla for the past couple of years, Intel has announced that they have hired Keller to lead their silicon engineering efforts.

After rumors on the matter overnight, in a press release that has gone out this morning, Intel confirmed that they have hired Jim Keller as a Senior Vice President. There, Keller will be heading up the 800lb gorilla’s silicon engineering group, with an emphasis on SoC development and integration. Beyond this, Intel’s press release is somewhat cryptic – especially as they tend not to be very forward about future processor developments. But it’s interesting to note that in a prepared statement included with the press release, Dr. Murthy Renduchintala – Intel’s Chief Engineering Officer – said that the company has “embarked on exciting initiatives to fundamentally change the way we build the silicon as we enter the world of heterogeneous process and architectures,” which may been seen as a hint of Intel’s future direction.

What is known for sure is that for most of the last decade, Keller’s engineering focus has been on low-power hardware. This includes not only his most recent stint at Tesla working on low voltage hardware, but also his time at Apple and PA Semiconductor developing Apple’s mobile SoCs, and even AMD’s Zen architecture is arguably a case of creating an efficient, low-power architecture that can also scale up to server CPU needs. So Keller’s experience would mesh well with any future development plans Intel has for developing low-power/high-efficiency hardware. Especially as even if Intel gets its fab development program fully back on track, there’s little reason to believe they’re going to be able to duplicate the manufacturing-derived performance gains they’ve reaped over the past decade.

As for any specific impact Keller might have on Intel’s efforts, that is a curiosity that remains to be seen. Keller’s credentials are second to none – he’s overseen a number of pivotal products – but it bears mentioning that modern processor engineering teams are massive groups working on development cycles that span nearly half a decade. A single rock star engineer may or may not be able to greatly influence an architecture, but at the same time I have to imagine that Intel has tapped Keller more for his leadership experience at this point. Especially as a company the size of Intel already has a number of good engineers at their disposal, and unlike Keller’s second run at AMD, the company isn’t recovering from a period of underfunding or trying to catch up to a market leader. In other words, I don’t expect that Intel is planning on a moment of Zen for Keller and his team.


One of Jim Keller’s Many Children: AMD’s Raven Ridge APU

Though with his shift to Intel, it’s interesting to note that Jim Keller has completed a de facto grand tour of the high performance consumer CPU world. In the last decade he’s worked for Apple, AMD, and now Intel, who are the three firms making the kind of modern ultra-wide high IPC CPU cores that we see topping our performance charts. Suffice it to say, there are very high-profile engineers of this caliber that these kind of companies will so openly court and/or attempt to pull away from the competition.

For those keeping count, this also marks the second high-profile architect from AMD to end up at Intel in the last 6 months. Towards the end of last year Intel picked up Raja Koduri to serve as their chief architect leading up their discrete GPU development efforts, and now Jim Keller is joining in a similar capacity (and identical SVP title) for Intel’s silicon engineering. Coincidentally, both Kodrui and Keller also worked at Apple for a time before moving to AMD, so while they haven’t been on identical paths – or even working on the same products – Keller’s move to Intel isn’t wholly surprising considering the two never seem to be apart for too long. So it will be exciting to see what Intel is doing with their engineering acquisitions over the coming years.

NVIDIA Releases 397.31 WHQL Drivers: Vulkan 1.1 Support, NVIDIA RTX, And More

NVIDIA Releases 397.31 WHQL Drivers: Vulkan 1.1 Support, NVIDIA RTX, And More

After quietly announcing legacy support status for Fermi GPUs and 32-bit operating systems earlier this month, NVIDIA today has released the first Release 396 drivers with 397.31 WHQL, headlined with Game Ready titles BattleTech and Frostpunk. As the first of a new driver branch, 397.31 is a wide-ranging update that introduces a number of new features, including NVIDIA RTX developer preview support, additional sampling rates for GPU HDMI audio, Vulkan 1.1 mainstream support, and tempfile auto-removal functionality for the display driver standalone installer. And as expected, Release 396 officially drops mainstream ‘Game Ready’ support for Fermi and 32-bit operating systems.

Release 396 Features for Consumers

Starting things off with driver installation itself, the display driver standalone installer now automatically removes the temporary files that it self-extracts at the beginning of the process. As for the less glamorous world of HDMI audio through video cards, the new NVIDIA HD audio driver version 1.3.37.1 now supports 32, 88.2, and 176.4 kHz sample rates for all GPUs, rounding out the existing driver support for 44.1, 48, 96, and 192 kHz. The newly-added sample rates are rather uncommon compared to 44.1 (e.g. music CDs) and 48 kHz (e.g. digital video formats), and even 96 and 192 kHz (e.g. professional music recording or audiophile), but still adds versatility to HTPC setups powered by discrete NVIDIA graphics. It appears that 20 bit sample size remains unsupported, and no change to codec support was mentioned.

Meanwhile, although full support for Vulkan 1.1 was made available in NVIDIA developer drivers when the API update was launched last month, 397.31 has brought it to the mainstream Game Ready branches as opposed to beta Windows drivers. For end-users, this is another step to running Vulkan 1.1 applications, though NVIDIA notes that 397.31 passes the Vulkan Conformance Test Suite version 1.1.0.3, as opposed to the major 1.1.1.0 update that was released earlier in April.

Last but not least, alongside with the usual Game Ready support for BattleTech and Frostpunk, both released yesterday, NVIDIA has also added or updated SLI profiles for Descenders, Frostpunk, Warhammer Vermintide 2, and Far Cry 5.

Release 396 Features for Developers: RTX, Optimus DLLs, & NVDECODE API Updates

A little over a month ago, Microsoft announced DirectX Raytracing (DXR), bringing a standardized API for hardware and software accelerated ray tracing under DX12. Alongside this, NVIDIA unveiled their RTX Technology, NVIDIA’s underlying runtime for DXR and other real-time raytracing solutions. RTX supports Volta and later GPUs to accelerate ray tracing via a combination of hardware and software. With today’s driver release, NVIDIA RTX has finally arrived in the form of a developer preview. As NVIDIA notes, developing RTX-accelerated DXR applications requires the following:

  • NVIDIA Volta GPU
  • NVIDIA driver version 396 or higher
  • Windows 10 RS4 (colloquially known as “Spring Creators Update”)
  • Microsoft DXR developer package (includes DXR-enabled D3D runtimes, HLSL compiler, and headers)

As for mainstream developers, the recent announcement of the Quadro V100 now brings a suitable Volta GPU solution with professional Quadro drivers outside of the deep-learning/HPC oriented Tesla Volta lineup and the prosumer Titan V.

Also of interest to developers is an update to the NVDECODE API, adding client capabilities to reconfigure decoder parameters without destruction and recreation of the decoder instance. These parameters include post-processing elements such as display resolution, cropping rectangle, and aspect ratio of the decoded frame. In practical terms, this benefits developers in situations where initializing the decoder instance would greatly add to the overall decode execution time; NVIDIA mentions an example of decoding multiple short clips of varying resolutions back-to-back.

* NVIDIA’s footnotes: *Diagram represents support for Pascal GPU family; **4:2:2 is not natively supported on HW; ***Support is codec dependent

This functionality will feature in NVIDIA’s Video Codec SDK 8.2, which is expected to release later this quarter. And though the 397.31 release notes refer to NVDECODE as a whole new API, it has been around since the release of the Video Codec SDK 7.0 in 2016, where it was re-named from NVCUVID API. As part of the Video Codec SDK, the NVDECODE API deals with video decode acceleration using NVIDIA GPUs’ dedicated ‘NVDEC’ hardware decoder blocks, comparable to the video encoding/decoding hardware blocks of Intel (Quick Sync Video) and AMD (VCE/UVD).

And on a more minor note, NVIDIA has included two new libraries (nvdlist.dll and nvdlistx.dll) to support Optimus and MSHybrid notebooks, referring to the capability of graphics load switching between integrated and discrete graphics adapters to save power and increase battery life.

Bugfixes and Open Issues

As always, today’s driver release includes bugfixes and open issues. The following bugs are addressed and resolved in 397.31:

  • DOOM crashes on the GeForce GTX 1080 Ti due to the driver reverting to OpenGL 1.1 when HDR is enabled
  • Far Cry 5 crashes on the GeForce GTX 1060 after a few minutes of gameplay
  • NvfbcPluginWindow temporarily prevents Windows from shutting down after launching a Steam game
  • Driver TDR error may occur when using Firefox.
  • Rise of the Tomb Raider experiences flickering/corruption on the GeForce GTX 1060 when opening the in-game options.
  • With V-Sync and SLI both enabled in NVIDIA Control Panel, Diablo III freezes after switching windows (Alt + Tab) a few times

397.31 has also introduced several new bugs, and NVIDIA has documented the following new issues in Windows 10:

  • With HDR turned on, full-screen video playback on Microsoft Edge on an HDR display may cause corruption of the video and desktop
    • To recover, manually power-cycle the monitor
  • TDR errors may occur on the GeForce GTX 1080 Ti when changing resolutions in-game for Warhammer Vermintide 2 running in DX12 mode
    • As a workaround, use the DX11 option
  • NVIDIA Surround hotkeys as set in NVIDIA Control Panel do not work
  • GeForce Experience’s “In-Game Overlay” option cannot be enabled, nor does ShadowPlay recording work
    • The Microsoft Media Foundation library must be installed in order to use these features

Release 396 also does not support Microsoft Surface Books, with support due in a future driver release. Naturally, 397.31 does not apply to Fermi GPUs and 32-bit OSes, nor will any future Game Ready drivers.

The updated drivers are available through the GeForce Experience Drivers tab or online at the NVIDIA driver download page. The latest GeForce Experience client can also be found separately on its own download page. More information on this update and further issues can be found in the 397.31 release notes.

Related Reading

Microsoft Rolls Out Xbox April Update: FreeSync And More

Microsoft Rolls Out Xbox April Update: FreeSync And More

Although we’ve not yet gotten an official launch date for the latest version of Windows 10, Microsoft’s Xbox team has once again beat them to the punch with a new version of the console OS called the April Xbox Update. Apparently, they&rsq…

AMD Announces Q1 FY 2018 Results: Big Gains

AMD Announces Q1 FY 2018 Results: Big Gains

This afternoon, AMD announced their earnings for the first quarter of their 2018 fiscal year, and their recent trend of performance has continued, with revenues for the quarter up 40% to $1.65 billion. Even more importantly, they’ve finally gotten to the point where they are able to achieve solid margins, which were 36% for this quarter, up 4% from last year. This led to operating incomes of $120 million, compared to just $11 million a year ago. Net income was $81 million for the quarter, compared to a $33 million loss last year. Earnings per share were $0.08, up from a $0.04 loss per share last year.

AMD Q1 2018 Financial Results (GAAP)
  Q1’2018 Q4’2017 Q1’2017
Revenue $1647M $1340M $1178M
Gross Margin 36% 34% 32%
Operating Income +$120M -$2M +$11M
Net Income +$81M -$19M -$33M
Earnings Per Share +$0.08 -$0.02 -$0.04

AMD has been in a situation where they’ve been relying on Non-GAAP measures to show the underlying business in the midst of restructuring debt and losses due to changing their wafer agreement, but these results are all GAAP, and all good. The big gains are thanks to the Computing and Graphics segment, which almost doubled in revenue compared to the same quarter last year. This group had revenues of $1115 million, compared to $573 million a year ago, which is a 94.6% increase. There’s likely no surprise here, but that increase is thanks to strong sales of both Radeon graphics, and Ryzen processors. Processor average selling price (ASP) increased thanks to more sales of Ryzen, and GPU ASP also increased, thanks to Vega, and likely the cryptocurrency craze which has definitely driven up prices. Operating income for this segment was $138 million for the quarter, compared to an operating loss of $21 million a year ago. Basically, there’s nothing but good news here, after far too long of struggling in this segment.

AMD Q1 2018 Computing and Graphics
  Q1’2018 Q4’2017 Q1’2017
Revenue $1115M $908M $573M
Operating Income +$138M +$33M -$21M

AMD’s other segment is their Enterprise, Embedded, and Semi-Custom segment, and this group is the ones that really carried them through the lean years, with AMD diversifying quite a bit into semi-custom SoCs, and it certainly helped that they got design wins in both the Sony PlayStation 4, and the Microsoft Xbox One. Revenue for this segment wasn’t quite as rosy as the Computing and Graphics, with quarterly revenues of $532 million, which was down 12% compared to Q1 2017. The revenue decline is attributed to lower semi-custom revenue, but somewhat offset by higher server and embedded revenue. EPYC processor revenue has helped, but not enough to offset the loss of revenue in semi-custom. Operating income for this segment was $14 million, compared to $55 million a year ago, but AMD attributed the majority of the operating income decrease to a licensing gain in Q1 2017 which inflated the numbers.

AMD Q1 2018 Enterprise, Embedded, and Semi-Custom
  Q1’2018 Q4’2017 Q1’2017
Revenue $532M $432M $605M
Operating Income +$14M -$13M +$55M

All Other had an operating loss of $32 million, compared to a loss of $23 million a year ago.

Overall, there’s little to be disappointed with here. Likely AMD is hoping for stronger EPYC sales to improve its segment, but the Computing and Graphics segment was strong enough to carry the day easily. Looking forward to next quarter, AMD is expecting revenues of $1.725 billion, plus or minus $50 million, which would be a 50% increase from Q2 2017.