SoylentNews is people
SoylentNews
The partnership will see Intel license SiFive's IP to create its own SiFive P550-based 64-bit SoC that it will fabricate on its new 7 nm node. It will form the basis of a new development platform Intel is calling Horse Creek, and will be made available to customers interested in exploring its potential in various applications involving embedded SoC tech. This could mean smartphones, but also cars, IoT products and the like. If Intel gets enough interest, it could take the relationship further. Intel hasn't yet revealed the technical specifications of the SoC, so we don't know whether it will be a single-core or multi-core platform, although the latter is likely. It's GPU tech is also unknown at this time, but Xe-based graphics are likely.
While the first Horse Creek SoCs will be ready next year, it isn't likely we will see any Intel RISC-V-based chips in commercially available products until 2023 at least.
SiFive recently announced two new high-performance 64-bit RISC-V cores, the Performance P550 and Performance P270:
SiFive compares the Performance P550 core to Arm's Cortex-A75 with higher performance in SPECint2006 and SPECfp2006 integer/floating-point benchmark, all [in] a much smaller area which would enable a quad-core P550 cluster on about the same footprint as a single Cortex-A75 core.
See also: Ubuntu 20.04/21.04 64-bit RISC-V released for QEMU, HiFive boards
Previously: Intel May Attempt to Acquire SiFive for $2 Billion
(Score: 2) by DannyB on Monday June 28 2021, @05:50PM (7 children)
I don't know if these numbers are exactly right. I am recalling them (hopefully correctly) from some lecture / presentation on YouTube.
Intel processor programming book is 1600 pages. And doesn't cover some subsets of instructions.
RISC-V instruction set fits on one page.
Gee that comparison sounds like Common Lisp vs Scheme. Scheme is(was) described (originally) in under 50 pages. In CLtL, even the index can't fit into 50 pages.
Is it time for the layers and layers of baggage and cruft in Intel's product line to finally be swept into the dustbin of history? How many transistors does it take to implement all that in hardware? They have a RISC like microcode, and the CISC instructions are converted into these micro ops in hardware?
What if you could use RSIC but the same fabrication technology and dedicate a lot of those transistors to more parallelism? More speculative execution? Even just more cores exposed to the system designer.
What if a system's CPU had so many cores that GPUs were unnecessary?
If you think a fertilized egg is a child but an immigrant child is not, please don't pretend your concerns are religious
(Score: 0) by Anonymous Coward on Monday June 28 2021, @06:04PM (1 child)
>> More speculative execution?
I think we know where that would lead...
(Score: 2) by DannyB on Monday June 28 2021, @09:32PM
Maybe that could be fixed by adding more management engines to the processor.
If you think a fertilized egg is a child but an immigrant child is not, please don't pretend your concerns are religious
(Score: 1, Insightful) by Anonymous Coward on Monday June 28 2021, @07:17PM (3 children)
Both Intel and AMD are well aware that x86 will be obsoleted eventually, that's why they are buying FPGA designers... Cutting-edge research and supercomputers can have specialized hardware processors, the rest of us will be just fine with processors that can optimize themselves for any workload with a few kernel calls, and run any legacy software with ease. I'd be more worried about VMWare than Intel...
(Score: 3, Interesting) by takyon on Monday June 28 2021, @07:35PM (2 children)
Both Intel and AMD are going to have "mainstream" 24-core CPUs and adopt variations of big.LITTLE within the next 3 years for better performance-per-area and power efficiency. x86 will be 10 years away from dying for years to come.
[SIG] 10/28/2017: Soylent Upgrade v14 [soylentnews.org]
(Score: 3, Interesting) by Mojibake Tengu on Monday June 28 2021, @07:59PM
It's not "x86", x86 is already dead. More and more distros abandoning 32-bit. It's AMD64, since the very first Athlon64 CPU when Intel had to license the AMD ISA to survive into the new millenium.
At one moment in history, we'll get pure AMD64 architecture, the CPU powered starting in full 64long addressing mode without all that 16-bit and 32-bit junk shit kept along only because of Microsoft, without all those instructions unusable and forbidden in 64long.
I remember a tiny piece of source code released for public in early 2000's, written by someone at AMD to demonstrate lockless data structures in pure assembly, the then new architecture was called X64 in headers, internally.
The edge of 太玄 cannot be defined, for it is beyond every aspect of design
(Score: 3, Insightful) by DannyB on Monday June 28 2021, @09:34PM
I think due to inertia, the x86 and x64 platform will be here for a very, very long time.
Heck, look at IBM mainframes. COBOL.
But I think a shift is coming where it won't be the cool shiny thing any more. It will be a way to run legacy software. These kinds of changes do not ever happen overnight.
If you think a fertilized egg is a child but an immigrant child is not, please don't pretend your concerns are religious
(Score: 0) by Anonymous Coward on Tuesday June 29 2021, @02:58PM
scheme isn't lisp...