Language Selection

English French German Italian Portuguese Spanish

AMD; Down On It's Luck Again

Filed under
Hardware

Those of us aware of AMD's tremendous potential as a chipmaker have to be disappointed by Apple's recent news that it has chosen Intel as the sole chip provider for its Macintosh computers. In addition to the Apple news, AMD continues to strike out many of the world's largest PC OEMs, such as Dell. If the Apple news is true, this would come as a yet another financial hit to AMD and its strategies. Unfortunately for AMD, it is trapped in a vicious circle of supply and demand.

While manufacturers like Dell, Apple and others do care about performance, value and possible alternative chip suppliers, they always have a difficult time picking AMD as a supplier. The reason is clear: limited supply. Larger OEMs like Dell (and even Apple) demand high volumes of chips to fulfill its orders. Since AMD can't provide such quantities, it's hindering its own success in the market. We can't exactly blame AMD with this because it has to work with the limited financial resources it has.

The vicious cycle for AMD operates somewhat like this: AMD wants to expand à AMD can't expand due to limited revenue à AMD proposes OEMs to offer PCs equipped with its chips à OEMs decline due to lack of enough chips à AMD is back to square one. Clearly, this is neither AMD's nor the OEM's fault. AMD wants to supply its chips, but OEMs need enough quantity to cover the millions of PCs they retail annually.

Since Intel's business ventures go beyond making processors, it can cope with the costs of opening new fabrication facilities, thereby increasing production and offering better deals to OEMs that need to purchase quantities in the millions. Additionally, Intel can accept some losses in its chip business here and there, while AMD really can't.

Though OEMs, including Apple, are aware of the architectural issues with Intel's microprocessors, it's not much of a concern to them. The majority of the market they retail to isn't aware of the issues, and the systems are "fast enough" for the customers they are targeting. How much of the general computing population do you think will know the performance difference between dual-core Intel and AMD microprocessors? Yes, that's right - not very many. The main thing many customers are looking for is affordability and brand recognition, and Intel can definitely offer that.

Full Story.

More in Tux Machines

PlayOnLinux For Easier Use Of Wine

PlayOnLinux is a free program that helps to install, run, and manage Windows software on Linux. It can also manage virtual C: drives (known as Wine prefixes), and download and install certain Windows libraries for getting some software to run on Wine properly. Creating different drives using different Wine versions is also possible. It is very handy because what runs well in one version may not run as well (if at all) on a newer version. There is PlayOnMac for macOS and PlayOnBSD for FreeBSD. Read
more

Linux Kernel: KPTI, SEV, CBS

  • Experimental KPTI Support For x86 32-bit Linux
    For the Kernel Page Table Isolation (KPTI) support currently within the Linux kernel for addressing the Meltdown CPU vulnerability it's currently limited to 64-bit on the x86 side, but for the unfortunate souls still running x86 32-bit operating systems, SUSE is working on such support.
  • AMD Secure Encrypted Virtualization Is Ready To Roll With Linux 4.16
    With the Linux 4.16 kernel cycle that is expected to begin immediately following the Linux 4.15 kernel debut on Sunday, AMD's Secure Encrypted Virtualization (SEV) technology supported by their new EPYC processors will be mainline. Going back to the end of 2016 have been Linux patches for Secure Encrypted Virtualization while with Linux 4.16 it will finally be part of the mainline kernel and supported with KVM (Kernel-based Virtual Machine) virtualization.
  • Deadline scheduler part 2 — details and usage
    Linux’s deadline scheduler is a global early deadline first scheduler for sporadic tasks with constrained deadlines. These terms were defined in the first part of this series. In this installment, the details of the Linux deadline scheduler and how it can be used will be examined. The deadline scheduler prioritizes the tasks according to the task’s job deadline: the earliest absolute deadline first. For a system with M processors, the M earliest deadline jobs will be selected to run on the M processors. The Linux deadline scheduler also implements the constant bandwidth server (CBS) algorithm, which is a resource-reservation protocol. CBS is used to guarantee that each task will receive its full run time during every period. At every activation of a task, the CBS replenishes the task’s run time. As the job runs, it consumes that time; if the task runs out, it will be throttled and descheduled. In this case, the task will be able to run only after the next replenishment at the beginning of the next period. Therefore, CBS is used to both guarantee each task’s CPU time based on its timing requirements and to prevent a misbehaving task from running for more than its run time and causing problems to other jobs.

Graphics: Mesa and AMDGPU

  • Mesa 17.3.3 Released With RADV & ANV Vulkan Driver Fixes
    Mesa 17.3.3 is now available as the latest point release for the Mesa 17.3 stable series. This bi-weekly point release to Mesa presents several RADV Vega/GFX9 fixes, various Intel ANV Vulkan driver fixes, a DRI3 fix, and random fixes to the OpenGL drivers like RadeonSI, Etnaviv, and even Swrast.
  • R600g "Soft" FP64 Shows Signs Of Life, Enabling Older GPUs To Have OpenGL 4 In 2018
    Most pre-GCN AMD graphics cards are still limited to OpenGL 3.3 support at this time due to not supporting FP64. Only the HD 5800/6900 series on R600g currently have real double-precision floating-point support working right now so at present they are on OpenGL 4.3 rather than 3.3, but those other generations may be catching up soon thanks to the "soft" FP64 code.
  • AMDGPU DC Gets More Raven Ridge Improvements, Audio Fixes
    Harry Wentland of AMD has sent out the latest batch of patches for the AMDGPU DC display code stack. Fortunately it lightens up the DRM driver by about six thousand lines thanks to removing some unused code. Besides gutting out a chunk of unused code, the DC code has a few audio fixes (no word yet on supporting newer audio formats with DC), fixes on driver unload, a "bunch" of continued Raven Ridge display updates, and various other code clean-ups.
  • AMDGPU Firmware Blobs Updated For Video Encode/Decode
    There are updated AMDGPU microcode/firmware files now available for recent Radeon GPUs. The updated firmware files now available via the main linux-firmware.git repository are centered around the video blocks: UVD video decoding, VCE video encode, and the new VCN video encode/decode block with Raven Ridge.

Games: DRAG, Geneshift, Balloonatics and More