Post by Admin on May 8, 2016 12:11:44 GMT -5
The present of the data center is owned by Chip-zilla and the x-86 hegemony.
The future will see Open Power (IBM's PowerPC chips) attack the high-end of that market, and ARM attack the low-end (middle?).
seekingalpha.com/article/3964295-google-ibm-arm-amd-begin-assault-intels-server-franchise?page=2
Exactly how the middle gets divided up is anyone's guess.
I would estimate that how the data center cpu market segments itself over the next few years is x-86 will shrink from almost total dominance to around 55%, ARM will rise from almost nothing to about 25%, Power PC will claim 15% (mostly from the high end), and 5% will be other assorted specialty chips. I think that there's a very good chance that INTC without total market dominance would have to sell its fabrication business and adopt a business model similar to ARM or at least a mostly "fab-less" model in order to compete without their present monopolistic lock on the server market.
It is still unclear how the ARM market specifically will shake-out. There are still around a half dozen competitors in various states of development. AMCC is a first-mover, and that's a good thing. What worries me is that competitors (yes competitors), like CAVM have realized that ARM will likely not compete at the very high-end of the market and that the lack of "brawny cores" is not the issue for most servers in the data centers.
www.nextplatform.com/2016/04/28/arm-server-chips-xeon-class-matter/
From the article above the point is made (below) that the vast majority of server cpu apps do NOT need brawny cores:
""These nits may not matter as much in the future as they appear to today. While public clouds require excellent per-core and per-socket performance to be able to handle the wide variety of workloads run by their enterprise customers, high-end processors are overkill for many of the fastest growing workloads. There may be a large opportunity for smaller, more efficient processors that are “right-sized” to match the needs of the IoT and other segments."
"A few examples may help illustrate the point. Facebook’s massive server farms do not require absolute screaming performance, and depend instead on lower-clocked and lower-cored Xeons, and more recently the new Xeon D in the “Yosemite” microserver, which can deliver sufficient single-socket performance and enough DRAM at lower prices and lower power envelopes. (See our recent coverage from the Open Compute Summit on Facebook’s server configurations.)"
"Similarly, content delivery networks do not require brawny cores, and this market will experience hyper-growth as virtual reality games begin shipping in earnest and as video continues to displace old fashioned photos in advertising and communications. Content deliver networks need good I/O and networking bandwidth, but they don’t require a lot of computational performance to pick up data off a disk drive and put it on the (right) virtual wire. Finally, in computationally intensive applications like deep learning, where the compute is all on a GPU or other accelerator, the CPU is such a small part of the equation that a more modest CPU like an ARM server chip could deliver better economics.""
CAVM has used the above fact to enter the ARM server market quickly, while still addressing the vast majority of that market, leaving the high-end to others. The actual relevant improvements may lie in the inter and intra chip interconnects, and not in the "brawny core". The high-speed interconnect is AMCC's specialty. Has AMCC gone outside of its circle of competence at potentially great expense (time and money)?
Perhaps Gopi is more of a visionary than he gets credit for. Only time will tell.
The future will see Open Power (IBM's PowerPC chips) attack the high-end of that market, and ARM attack the low-end (middle?).
seekingalpha.com/article/3964295-google-ibm-arm-amd-begin-assault-intels-server-franchise?page=2
Exactly how the middle gets divided up is anyone's guess.
I would estimate that how the data center cpu market segments itself over the next few years is x-86 will shrink from almost total dominance to around 55%, ARM will rise from almost nothing to about 25%, Power PC will claim 15% (mostly from the high end), and 5% will be other assorted specialty chips. I think that there's a very good chance that INTC without total market dominance would have to sell its fabrication business and adopt a business model similar to ARM or at least a mostly "fab-less" model in order to compete without their present monopolistic lock on the server market.
It is still unclear how the ARM market specifically will shake-out. There are still around a half dozen competitors in various states of development. AMCC is a first-mover, and that's a good thing. What worries me is that competitors (yes competitors), like CAVM have realized that ARM will likely not compete at the very high-end of the market and that the lack of "brawny cores" is not the issue for most servers in the data centers.
www.nextplatform.com/2016/04/28/arm-server-chips-xeon-class-matter/
From the article above the point is made (below) that the vast majority of server cpu apps do NOT need brawny cores:
""These nits may not matter as much in the future as they appear to today. While public clouds require excellent per-core and per-socket performance to be able to handle the wide variety of workloads run by their enterprise customers, high-end processors are overkill for many of the fastest growing workloads. There may be a large opportunity for smaller, more efficient processors that are “right-sized” to match the needs of the IoT and other segments."
"A few examples may help illustrate the point. Facebook’s massive server farms do not require absolute screaming performance, and depend instead on lower-clocked and lower-cored Xeons, and more recently the new Xeon D in the “Yosemite” microserver, which can deliver sufficient single-socket performance and enough DRAM at lower prices and lower power envelopes. (See our recent coverage from the Open Compute Summit on Facebook’s server configurations.)"
"Similarly, content delivery networks do not require brawny cores, and this market will experience hyper-growth as virtual reality games begin shipping in earnest and as video continues to displace old fashioned photos in advertising and communications. Content deliver networks need good I/O and networking bandwidth, but they don’t require a lot of computational performance to pick up data off a disk drive and put it on the (right) virtual wire. Finally, in computationally intensive applications like deep learning, where the compute is all on a GPU or other accelerator, the CPU is such a small part of the equation that a more modest CPU like an ARM server chip could deliver better economics.""
CAVM has used the above fact to enter the ARM server market quickly, while still addressing the vast majority of that market, leaving the high-end to others. The actual relevant improvements may lie in the inter and intra chip interconnects, and not in the "brawny core". The high-speed interconnect is AMCC's specialty. Has AMCC gone outside of its circle of competence at potentially great expense (time and money)?
Perhaps Gopi is more of a visionary than he gets credit for. Only time will tell.