hi,
here is the summary of the various responses I recd to the Ross Hypersparc qn:
The original posting was :
>HI,
>We are on the verge of purchasing Ross's 90 MHz hypersparc for our new SS-20.
>I have read some articles on the company itself( a sub. of Fujitsu) and perfor->mance statistics. Have also read about the hypersparc architecture.
>
>My qn is : has anyone been using these chips and can they offer an opinion?
>Many thanks,
>Vikram
The answers:
>From molly!cable.lvsun.COM!carl@uunet.uu.net Mon Feb 27 19:31:26 1995
Return-Path: <molly!cable.lvsun.COM!carl@uunet.uu.net>
Received: by demosun10.nyc.future.com (The_Future_Now.080594); Mon, 27 Feb 95 19:31:23 EST; id AA06043
Received: from uucp1.UU.NET by relay3.UU.NET with SMTP
id QQyeyr02082; Mon, 27 Feb 1995 19:27:00 -0500
Received: from lvsun.UUCP by uucp1.UU.NET with UUCP/RMAIL
; Mon, 27 Feb 1995 19:26:36 -0500
Received: from cable.noname by lvsun.COM (4.1/SMI-4.1)
id AA24825; Mon, 27 Feb 95 16:20:10 PST
Received: by cable.noname (4.1/SMI-4.1)
id AA06281; Mon, 27 Feb 95 16:20:09 PST
Date: Mon, 27 Feb 95 16:20:09 PST
From: carl@cable.lvsun.COM (carl shapiro x4655)
Message-Id: <9502280020.AA06281@cable.noname>
To: khullar@nyc.future.com
Subject: Re: Ross Hypersparc -- Any experiences
Newsgroups: comp.sys.sun.admin
In-Reply-To: <3itd2n$2cd@demosun10.nyc.future.com>
Organization: Las Vegas Sun
Cc:
Status: R
In article <3itd2n$2cd@demosun10.nyc.future.com> you write:
>HI,
>We are on the verge of purchasing Ross's 90 MHz hypersparc for our new SS-20.
>I have read some articles on the company itself( a sub. of Fujitsu) and perfor-
>mance statistics. Have also read about the hypersparc architecture.
>
>My qn is : has anyone been using these chips and can they offer an opinion?
>Many thanks,
>Vikram
>
We've had *very* good luck with a pair of Ross 72MHz chips
in an SS-10, running Ross's supplied patches to SunOS 4.1.3.
The system runs 24x7, with load ranging from 2-30 users, no
problems of any kind since installation in August '94.
-- Carl
>From jayk@mail.eskimo.com Mon Feb 27 20:41:38 1995
Return-Path: <jayk@mail.eskimo.com>
Received: by demosun10.nyc.future.com (The_Future_Now.080594); Mon, 27 Feb 95 20:41:35 EST; id AA07153
Received: by mail.eskimo.com (5.65c/1.35)
id AA19903; Mon, 27 Feb 1995 17:34:16 -0800
Date: Mon, 27 Feb 1995 17:34:16 -0800
From: jayk@mail.eskimo.com (Jay Koski)
Message-Id: <199502280134.AA19903@mail.eskimo.com>
To: khullar@nyc.future.com (Vikram Khullar)
Subject: Re: Ross Hypersparc -- Any experiences
Newsgroups: comp.sys.sun.admin
Organization: Eskimo North (206) For-Ever
X-Newsreader: TIN [version 1.2 PL2]
Status: R
In article <3itd2n$2cd@demosun10.nyc.future.com> you wrote:
: HI,
: We are on the verge of purchasing Ross's 90 MHz hypersparc for our new SS-20.
: I have read some articles on the company itself( a sub. of Fujitsu) and perfor-
: mance statistics. Have also read about the hypersparc architecture.
: My qn is : has anyone been using these chips and can they offer an opinion?
: Many thanks,
: Vikram
We run the quad 90 Ross CPUs, they are very nice. You may want to
email nanook@eskimo.com for more information regarging the CPUs.
>From nanook@eskimo.com Mon Feb 27 23:23:18 1995
Return-Path: <nanook@eskimo.com>
Received: by demosun10.nyc.future.com (The_Future_Now.080594); Mon, 27 Feb 95 23:23:10 EST; id AA09265
Received: from eskimo.com by mx.smtp.psi.net (8.6.9/SMI-4.1.3-PSI)
id WAA29787; Mon, 27 Feb 1995 22:29:49 -0500
Received: by eskimo.com (5.65c/1.35)
id AA19507; Mon, 27 Feb 1995 19:28:20 -0800
Date: Mon, 27 Feb 1995 19:28:20 -0800
From: nanook@eskimo.com (Robert Dinse)
Message-Id: <199502280328.AA19507@eskimo.com>
To: khullar@nyc.future.com
Subject: Re: Ross Hypersparc -- Any experiences
News-Path: openwx!news.sprintlink.net!howland.reston.ans.net!math.ohio-state.edu!uwm.edu!news.alpha.net!news.mathworks.com!uunet!in1.uu.net!psinntp!demosun10.nyc.future.com!khullar
References: <3itd2n$2cd@demosun10.nyc.future.com>
Status: RO
We're running Ross quad 90's on a 4/670MP platform. The performance
increase from the 40mhz Ross CPU's it came with roughly scaled with the CPU
speed.
Initially I had some problems, a memory card was had the speed selection
for the SIMMS set to 100ns when they were 80ns SIMMS, this did not cause any
noticable problems at 40mhz, but at 90mhz it made it totally non-functional.
Once that was resolved they have been very stable, this under 4.1.3_U1 with the
Hypersparc patches from Ross.
Getting them from Ross was another matter, it took 4 months from order to
delivery and this only after a lot of angry letters.
>From cbarnard@cs.uchicago.edu Tue Feb 28 14:36:01 1995
Received: by demosun10.nyc.future.com (The_Future_Now.080594); Tue, 28 Feb 95 14:35:58 EST; id AA22566
From: cbarnard@cs.uchicago.edu
Return-Path: <cbarnard@cs.uchicago.edu>
Received: by salmon.uchicago.edu (4.1/UofC4.4x)
id AA01700; Tue, 28 Feb 95 13:30:26 CST
Date: Tue, 28 Feb 95 13:30:26 CST
Message-Id: <9502281930.AA01700@salmon.uchicago.edu>
To: khullar@nyc.future.com (Vikram Khullar)
In-Reply-To: khullar@nyc.future.com's message of 27 Feb 1995 20:35:35 GMT
Subject: Ross Hypersparc -- Any experiences
Status: R
From: khullar@nyc.future.com (Vikram Khullar)
Newsgroups: comp.sys.sun.admin
Date: 27 Feb 1995 20:35:35 GMT
We are on the verge of purchasing Ross's 90 MHz hypersparc for our new SS-20.
I have read some articles on the company itself( a sub. of Fujitsu) and perfor-
mance statistics. Have also read about the hypersparc architecture.
My qn is : has anyone been using these chips and can they offer an opinion?
We have quad 55Mhz Ross processors installed in two of our SparcStation 10s
and a dual 90Mhz Ross processor installed in a SparcStation 20. We've been
very happy with both. The speedup isn't linear (ie, the 55Mhz isn't just
under twice as fast as the 30Mhz as we expected) but there is definitely
an improvement. Be sure to buy more memory when you get the additional
processors, btw.
Christopher L. Barnard http://cs-www.uchicago.edu/~cbarnard
Acting Director, Computer Systems cbarnard@cs.uchicago.edu
Computer Science Department, University of Chicago (312) 702-8850
PGP public key available via finger or the MIT PGP keyserver .
>From p45dpmg@sherritt.ca Tue Feb 28 14:40:21 1995
Return-Path: <p45dpmg@sherritt.ca>
Received: by demosun10.nyc.future.com (The_Future_Now.080594); Tue, 28 Feb 95 14:40:17 EST; id AA22696
Received: from turin.ucs.ualberta.ca by bock.ucs.ualberta.ca with SMTP
(8.6.5/UA2.0.0.93Dec20) id MAA11186
for <khullar@nyc.future.com>; Tue, 28 Feb 1995 12:35:53 -0700
Received: from sherritt.ca (isglcr.sherritt.CA) by turin.ucs.ualberta.ca (4.1)
id AA26366; Tue, 28 Feb 95 12:35:42 MST
Received: from bigbrd.sherritt.ca by sherritt.ca (4.1/SMI-4.1)
id AA18236; Tue, 28 Feb 95 12:43:07 MST
Received: by bigbrd.sherritt.ca (4.1/SMI-4.1)
id AA11555; Tue, 28 Feb 95 12:34:49 MST
Date: Tue, 28 Feb 95 12:34:49 MST
From: p45dpmg@sherritt.ca (Peter Graw)
Message-Id: <9502281934.AA11555@bigbrd.sherritt.ca>
To: khullar@nyc.future.com
Subject: Re: Ross Hypersparc -- Any experiences
Newsgroups: comp.sys.sun.admin
In-Reply-To: <3itd2n$2cd@demosun10.nyc.future.com>
Organization: /usr/lib/news/organi[sz]ation
Cc:
Status: R
We're running 4 55MHz modules in a Sparc 10. A few Sun OS 4.1.3 patches, and
off we went.
-- -------------------------------------------------- Peter Graw (p45dpmg@sherritt.ca) Sherritt Inc. Information Technology>From negativl@netcom.com Tue Feb 28 19:30:58 1995 Return-Path: <negativl@netcom.com> Received: by demosun10.nyc.future.com (The_Future_Now.080594); Tue, 28 Feb 95 19:30:54 EST; id AA26620 Received: by netcom8.netcom.com (8.6.9/Netcom) id QAA11916; Tue, 28 Feb 1995 16:27:06 -0800 Date: Tue, 28 Feb 1995 16:27:06 -0800 From: negativl@netcom.com (Raymond Wong) Message-Id: <199503010027.QAA11916@netcom8.netcom.com> To: khullar@nyc.future.com Subject: Re: Roos Hypersparc on Sol2.x Newsgroups: comp.sys.sun.admin In-Reply-To: <3ivkfp$jtt@demosun10.nyc.future.com> Organization: NETCOM On-line Communication Services (408 261-4700 guest) Cc: Status: R
WE're running a SS20 612 with the Ross 90 Mhz package, but this is the first 20 and the first 2.3 box we've ever bought, so I can't tell you how it compares to the normal 20-612. Installed perfectly, no changes to software needed, and everything seems to run OK, but like I said, I'm not sure what it would be normally.
In article <3ivkfp$jtt@demosun10.nyc.future.com> you write: >Hi, >Thanks for the responses I recd by emasil about Ross performance.It would be >nice if someone shares their Solaris 2.x experience with Ross. All the responses I >recd were pertaining to SunOS 4.x > >If there are any souls out there with that situation, >please let me know... >Vikram >
We are running Solaris 2.3 on about 40 machines (SS2, SS10, SS20, SC2000, 4/690, IPX, etc.). The SS10 and 4/6x0 machines are running with Ross HyperSPARC CPU modules. The SS10s have single-CPU 66MHz modules, and the 4/6x0 systems are running a mix of dual-CPU 66MHz and single-cpu 90MHz modules (not in the same machine, of course.) We've have no major problem, other than the occasional module showing up DOA, but Ross exchanged them right away. >From what we've observed, the dual 66MHz modules are faster at FP than the TI SuperSPARC @50MHz. The 90s are faster than that, but getting dual or quad 90s makes the biggest difference. Anything else you want to know? Scott -- Scott McClure (scottm@rain.org) | "The (Information) Superhighway? That System Test & Support Engineer | sounds like something long and boring Brandes Associates, Inc. | and kills 50,000 people each year" - D.C.
This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:10:17 CDT