Domain: sequent.com
Stories and comments across the archive that link to sequent.com.
Comments · 7
-
IBM Gaining Marketability in Mainframe Industry
IBM makes mainframes. They are third in the industry behind Sun and Compaq (DEC), respectively. I think they are aiming the 64-bit PPC Linux port to give them a boost in the market. Native linux has scalability problems but IBM has been putting an effort into developing large scale Linux solutions in recent projects. Combined with a recent deal with Redhat, they seem to be surging forwards into the linux mainframe market.
-
Re:36bit architectureI really don't think that for example Intel hardware is designed for mission-critical situations.
Well, your common or garden variety PC desktop or "server" (in most cases, a desktop PC with more drive bays, if you buy it from someone like Gateway) certainly isn't. But that's a function of the whole system, not the CPU, as Sequent show.
-
Re:Bandwidth...You're right on with the technical details, but I'm going to have to say that Cray is still Cray.
:) Ironically, Cray's new SV2 is ccNUMA,also. However, NUMA isn't an SGI- or Cray-only thing these days, IBM (Sequent) is also doing it.As an aside for those who don't care to read whitepapers and didn't already know this:
cc = Cache Coherent. The memory interface logic keeps track of what cache has what data cached so the data doesn't get corrupted.
NUMA = Non-Uniform Memory Access. There's no one point of access to core memory. Each node or board (or processor, let's just call them system blocks) has its own memory inteface, and if other system blocks need some data from another system block's memory, it must ask for it. This removes bottlenecks to relatively slow (insert-favorite-flavor-of-)DRAM. This allows for a greater theoretical system-bandwidth peak through parallelization of slow storage pipes. Writing 16 bytes, 1 byte at a time will always be 16 times slower than writing 16 bytes to 16 banks of DRAM all at once. -
Re:CC-NUMA & NUMA-QccNUMA (r) is SGI's cache coherent NUMA. There are a lot of sophisticated tricks played with memory, but it all boils down to a system with lower memory latency over the 'link than the old Power Series machines had locally. The coherent cache turns out to be handy, too. I never did delve deeply into those machines, but they definitely are kewl. And they scale like no one's business. More info can be found here.
NUMA-Q is the Sequent technology. It is also cache coherent according to this paper, but the details are lacking. It does not appear to scale as well as SGI's NUMA, though.
-
Re:It's still bus based
According to some of the white papers on Sequent's web page, off-quad (ie, non local) memory access are around 20 times slower than local access. In other words, although you can get away without changing your programming model in theory, in practice you have to.
-
Multiple OS-s
What strikes me as the bigest difference between NUMA-s and other multi-processor machines would be the idea of using different OS-s (well, actually only their Unix and M$ NT) on the multi-procesor system so that each OS gets one (or more?) processor to work with. At least that's what they claim here:
NUMACenter enables simultaneous use of both Unix and Windows NT on one system.
-
here's NUMA
oops.. further review of the Sequent site (and some delving into the old stuff, I think) led me to find this, which might help some. It seems to look technical, and I'll bet it's what I was looking for...
:)