mat@amdahl.uts.amdahl.com (Mike Taylor) (05/06/88)
Just a short posting from the mainframe world, since we have announced
a new processor family for delivery starting in June, the Amdahl 5990.
Since this machine supports Amdahl UTS, it is a Unix machine. We think
this is probably the fastest scalar machine (both integer and floating
point) around, so I've taken the liberty of posting a few details.
Model 5990-700 5990-1400
Number of CPUs 2 4
Main storage (MB) 64-256 128-512
Expanded storage(MB) 128-1024 128-2048
I/O Channels 32-64 64-128
The 5990 CPU operates at a cycle time of 10ns. It uses 180ps. 3K and
10K gate ECL logic. There are also combined logic and RAM chips
that have 1200 280 ps. logic gates and 16 kb RAM with 2.8 ns. access time.
Main storage is 55ns. 256 kb SRAM. Each CPU occupies a single board,
including 64KB instruction and operand caches. A six-stage pipeline
is capable of a peak rate of one instruction completion per cycle.
Each channel is capable of managing and queuing I/O for up to 256 devices
and supports data transfer rates up to 4.5 MB/sec.
RAS features include storage patrol, internal redundancy, history logic,
automatic scan out analysis (ASOA), failure analysis system (FAS) and
AMDAC remote maintenance.
The Multiple Domain Facility allows system resources to be allocated to
up to 4 (5990-700) or 7 (5990-1400) domains, which may independently
support different SCPs (UTS, MVS, VM) and different architectural features
(S/370, S/370-XA, ESA/370). Each domain can configure up to 4080
I/O devices.
--
Mike Taylor ...!{ihnp4,hplabs,amdcad,sun}!amdahl!mat
[ This may not reflect my opinion, let alone anyone else's. ]bcase@Apple.COM (Brian Case) (05/06/88)
In article <30872@amdahl.uts.amdahl.com> mat@amdahl.uts.amdahl.com (Mike Taylor) writes:
[Lots of good information about the 5990, impressive sounding!!!]
Good God man, enough of this drivel; dhrystones, damnit, I want dhrystones!
:-) :-) :-)
[From "Breaking Away:" "I want American food, damnit, I want French Frys."]chuck@amdahl.uts.amdahl.com (Charles Simmons) (05/08/88)
In article <9337@apple.Apple.Com> bcase@apple.UUCP (Brian Case) writes: >In article <30872@amdahl.uts.amdahl.com> mat@amdahl.uts.amdahl.com (Mike Taylor) writes: > > [Lots of good information about the 5990, impressive sounding!!!] > >Good God man, enough of this drivel; dhrystones, damnit, I want dhrystones! >:-) :-) :-) > >[From "Breaking Away:" "I want American food, damnit, I want French Frys."] We just got dhrystone 2.0 results hot off the iron. The bottom line is that we can do 74000 dhrystones using our current C compiler and 91000 dhrystones using our experimental GNU C compiler. And just in time too. MIPS was starting to get a little bit too close. -- Cheers, Chuck
brucek@hpsrla.HP.COM (Bruce Kleinman) (05/10/88)
+------- | ... 16 kb RAM with 2.8 ns. access time ... Main storage is 55ns. 256 kb SRAM. +------- This looks like a great way to avoid the DRAM shortage! +------- | Good God man, enough of this drivel; dhrystones, damnit, I want dhrystones! +------- Dhrystones? We don't need no stinkin' dhrystones! Let's see some really *useful* benchmark numbers ... like Puzzle and Sieve. Bruce.
mash@mips.COM (John Mashey) (05/10/88)
In article <31124@amdahl.uts.amdahl.com> chuck@amdahl.uts.amdahl.com (Charles Simmons) writes: .... >We just got dhrystone 2.0 results hot off the iron. The bottom line >is that we can do 74000 dhrystones using our current C compiler and >91000 dhrystones using our experimental GNU C compiler. > >And just in time too. MIPS was starting to get a little bit too close. Sigh. Not this year. (Like I said when "brash micros vs Big Iron" was the topic.) We'll just have to hang around while in the 25-45% range.... But then, none of our CPU performance graphs claimed crossover of VLSI RISC micro versus mainframe uniprocessor until 1989. :-) More seriously, thanx for posting the numbers: it's good to get some Big Iron numbers showing up in this newsgroup. -- -john mashey DISCLAIMER: <generic disclaimer, I speak for me only, etc> UUCP: {ames,decwrl,prls,pyramid}!mips!mash OR mash@mips.com DDD: 408-991-0253 or 408-720-1700, x253 USPS: MIPS Computer Systems, 930 E. Arques, Sunnyvale, CA 94086
mash@mips.UUCP (05/12/88)
In article <3460014@hpsrla.HP.COM> brucek@hpsrla.HP.COM (Bruce Kleinman) writes: >+------- >| Good God man, enough of this drivel; dhrystones, damnit, I want dhrystones! >+------- >Dhrystones? We don't need no stinkin' dhrystones! Let's see some really >*useful* benchmark numbers ... like Puzzle and Sieve. This is one more reminder that we really could use some more good, realistic, substantive, non-toy, public domain, commonly-cited integer benchmarks. It would be awfully nice to get at least 6-10 good ones that people agreed had some representative qualities. -- -john mashey DISCLAIMER: <generic disclaimer, I speak for me only, etc> UUCP: {ames,decwrl,prls,pyramid}!mips!mash OR mash@mips.com DDD: 408-991-0253 or 408-720-1700, x253 USPS: MIPS Computer Systems, 930 E. Arques, Sunnyvale, CA 94086
eugene@pioneer.UUCP (05/12/88)
In article <3460014@hpsrla.HP.COM> brucek@hpsrla.HP.COM (Bruce Kleinman) writes: >+------- >| Good God man, enough of this drivel; dhrystones, damnit, I want dhrystones! >+------- >Dhrystones? We don't need no stinkin' dhrystones! Let's see some really >*useful* benchmark numbers ... like Puzzle and Sieve. To keep the motion picture theme, remember the first line was a modification from the film "Breaking Away." See it again, and see the scene where he says: "Refund?! Refund?!. . ." Wouldn't you just love to be able to send some of your machines back? [Now back to the stinkin' "The Treasure of Sierra Madre."] Another gross generalization from --eugene miya, NASA Ames Research Center, eugene@aurora.arc.nasa.gov resident cynic at the Rock of Ages Home for Retired Hackers: "Mailers?! HA!", "If my mail does not reach you, please accept my apology." {uunet,hplabs,hao,ihnp4,decwrl,allegra,tektronix}!ames!aurora!eugene "Send mail, avoid follow-ups. If enough, I'll summarize."
chuck@amdahl.uts.amdahl.com (Charles Simmons) (05/14/88)
In article <2175@winchester.mips.COM> mash@winchester.UUCP (John Mashey) writes: >This is one more reminder that we really could use some more good, realistic, >substantive, non-toy, public domain, commonly-cited integer benchmarks. >It would be awfully nice to get at least 6-10 good ones that people >agreed had some representative qualities. >-- >-john mashey DISCLAIMER: <generic disclaimer, I speak for me only, etc> >UUCP: {ames,decwrl,prls,pyramid}!mips!mash OR mash@mips.com >DDD: 408-991-0253 or 408-720-1700, x253 >USPS: MIPS Computer Systems, 930 E. Arques, Sunnyvale, CA 94086 Possibly this is an inappropriate newsgroup for discussing benchmarks. If so, I'm sure all you kind people will send me mail so that I'll never talk about benchmarks here again. Since John has quite a bit of experience working with benchmarks, I'd be interested in hearing his thoughts on what makes a benchmark good, realistic, and substantive. Maybe other people would like to contribute their ideas as well? As a question, can a benchmark be good and useful if it runs in time proportional to N*N where N is an input parameter? For example, the drhystone benchmark runs in time c*N where N is the number of loops to execute. This makes it very easy to compare dhrystone results obtained from machines where a different number of loops were run on each different machine. For a more complex program like 'sieve', it might be difficult to obtain comparable results across a broad class of machines. Thanks, Chuck