Talk:Cray-3

Cray-3 has been listed as one of the Engineering and technology good articles under the good article criteria. If you can improve it further, please do so. If it no longer meets these criteria, you can reassess it.
Review: December 2, 2016. (Reviewed version).

A fact from Cray-3 appeared on Wikipedia's Main Page in the Did you know column on 3 January 2017 (check views). The text of the entry was as follows:

Did you know... that when Seymour Cray designed the Cray-3 supercomputer, he insisted that the processor fit into a 1-cubic-foot (0.028 m³) block?

A record of the entry may be seen at Wikipedia:Recent additions/2017/January. The nomination discussion and review may be seen at Template:Did you know nominations/Cray-3.

Wikipedia

What logic was implemented in GaAs

Did the Cray-3 use ECL or what logic ? Rod57 (talk) 17:31, 26 October 2008 (UTC)[reply]

GaAs. 198.123.51.6 (talk) 17:48, 17 March 2009 (UTC)[reply]

GaAs, sure, but as TTL, ECL, CMOS or what ? - Rod57 (talk) 05:22, 3 January 2016 (UTC)[reply]

I think GaAs logic circuits are usually nMOS-like. Dicklyon (talk) 06:09, 3 January 2016 (UTC)[reply]

Thanks. Actually I just found a GaAs source saying : Cray used Capacitor Diode FET logic (CDFL) ^[1] (It also says the slower Direct-coupled FET logic (DCFL) is similar to NMOS logic.) I've put more at Gallium_arsenide#GaAs_digital_logic - Rod57 (talk) 06:58, 3 January 2016 (UTC)[reply]

^ Dennis Fisher and I. J. Bahl (1995). Gallium Arsenide IC Applications Handbook. Vol. 1. Elsevier. p. 61. ISBN 978-0-12-257735-2. 'Clear search' to see pages

Good find. Dicklyon (talk) 07:12, 3 January 2016 (UTC)[reply]

Compatibility, NOT

The person who wrote the line about the Cray-3 being compatible with the Cray-2 has never run on any more than a single one of Seymour's machines if any. This is a common mistake by on lookers from other architectures. They have no clue about supercomputing. That line needs to be struck. The only thing which would have been vaguely comparable would be the OS. Source code would have had to have been recompiled into new binaries. 66.122.34.11 (talk) 00:47, 9 December 2011 (UTC)[reply]

You may be right. Do you know any books or manuals or magazines or other WP:RELIABLE sources that discuss the compatibility (or lack of compatibility) between the Cray-2 and the Cray-3?

So far, we have a reference in the article that says that, early in the design of the Cray-3, it was *intended* to be compatible with the Cray-2, but we don't yet have any sources that specifically say whether the actual shipped computers were binary compatible or source compatible or not. --DavidCary (talk) 23:53, 25 August 2015 (UTC)[reply]

I wrote the Fortran and C runtime libraries for the Cray-3 and the Cray-4 (and I took the photograph of Seymour and the Cray-3 as well as the standalone photo of of the Cray-3 that's on the page). The Cray-3 was ISA compatible with the Cray-2 (in fact, we did our SW development -- including the assembler code -- on the single quadrant Cray-2 snq2 in-house and Cray-2 SN2025 at NERSC). As part of the split-up Cray Computer acquired the rights to the Cray-2, although we never attempted to sell any of them. With 20/20 hindsight, we probably should have - we would have been in a better financial situation if we'd had an income stream. The Cray-4 was *not* ISA compatible with the Cray-2/Cray-3 (no local memory, an actual hardware exchange jump instruction, T registers, some different load/store instructions to use the additional memory paths, and some new vector instructions). By the way, I'd appreciate a credit line on the photographs. Wikipedia certainly has my permission to use them.

-- Steve Gombosi

While I'm at it: CSOS was not "based on Unicos". It was a clean Unix SVR4 port. We started from scratch on the compilers, as well.

-- Steve Gombosi

Removed statement

While citing up the article, I removed this statement:

In practice, two of the processors were removed and the machine was used unofficially for some time after that.

I clearly remember reading this, but can no longer find the source. I will re-insert it if I can. Maury Markowitz (talk) 16:28, 20 October 2016 (UTC)[reply]

GA Review

This review is transcluded from Talk:Cray-3/GA1. The edit link for this section can be used to add comments to the review.

Reviewer: Spinningspark (talk · contribs) 15:01, 16 November 2016 (UTC)[reply]

Looking... SpinningSpark 15:01, 16 November 2016 (UTC)[reply]

Lead

I feel Cray (the person) should be mentioned earlier in the lead and his relationship to the company made clear.

Indeed, I have cleaned all of this up a bit.

Background

The third paragraph has an ambiguity between "he" and "they" in the first and second sentence.

Fixed.

Development

"original Chippewa Falls lab". It is not clear what this is, presumably it is the CRI lab where the Cray 3 development began. This could be stated explicitly, perhaps introducing the name earlier in the article as well. Also, if this is Chippewa Falls, Wisconsin it should be wikilinked on first mention.

Fixed and fixed.

"with the company retaining 10% of the new company's stock". Might be better to explicitly say which company, ie "with CRI retaining..."

Fixed.

Cray...worked under contract. This did, however, require the lab to move to a new building" This seems confusing; it reads as if Cray working under contract caused the move to a new building. Is the move to a new building synonymous with the move from Chippenwa Falls? If so, a change of building is obvious. Why not just say the move caused further delays. Also, why is it "further", no previous delays were mentioned. One can read between the lines that the low priority delayed development, but if that is the case then it should be stated explicitly.

They did move twice - CRI still owned their original building in Colorado, when the company was spun off they had to leave. Should be clearer now.

"In service, the static RAM proved to be problematic, it was discovered that the square root code contained a bug..." How can a bug in the code be a problem with the RAM?

Fixed.

MPP is not defined or wikilinked.

Defined.

References

The external links tool shows several deadlinks and other problems.

Ok I'll work on these... fixed. Generally, what to do about dead urls in external links? Is that something there's an expectation on, or just remove them?

Overall, a nice, well presented article. Very close to GA standard and there'll be no problem passing once these few issues are fixed. SpinningSpark 17:05, 16 November 2016 (UTC)[reply]

Sorry for my tardy reply, I checked back a few times after you posted the original note on my page but then stopped checking! Maury Markowitz (talk) 14:11, 2 December 2016 (UTC)[reply]

Unfortunately, in the course of fixing problems you seem to have mangled this sentence: "Work started on the design in 1988 at [Cray Research]]'s (CRI), while other teams company were working on competing designs." It's ungrammatical, doesn't make sense, and CRI is not an acronym of "Cray Research's".
On the external links, a distinction should be made between links in the external links section and links in references. Deadlinks in the external links section are useless to readers and can be removed. For references, a further distinction should be made between links to online resources and convenience links of printed resources (journal articles, books etc). For online resources, the link should be retained as long as the text it is referencing remains in the article and a replacement has not been found. If possible, find a copy of the page in an archive (Wayback Machine, WebCite etc). Most citation templates have parameters that allow adding an archive link. See WP:DEADREF for more information. Convenience links, on the other hand, are not essential and can be removed or replaced if dead, but the reference itself should be retained. In any case, fixing dead links is not a GA requirement, so I only need you to fix the sentence above before I can pass it. SpinningSpark 18:39, 2 December 2016 (UTC)[reply]

OK I think I got all the dead links - which was basically all of them! - and fixed the issue you noted with my last edit. Let me know! Maury Markowitz (talk) 19:06, 2 December 2016 (UTC)[reply]

You still haven't said what the "I" in "CRI" stands for (presumably "Inc.") but I can live with that—passing article SpinningSpark 20:01, 2 December 2016 (UTC)[reply]

External links modified

Hello fellow Wikipedians,

I have just modified one external link on Cray-3. Please take a moment to review my edit. If you have any questions, or need the bot to ignore the links, or the page altogether, please visit this simple FaQ for additional information. I made the following changes:

Added archive https://web.archive.org/web/20080507080148/http://www.hoise.com:80/primeur/96/pr-96-oct/CL-PR-10-96-3.html to http://www.hoise.com/primeur/96/pr-96-oct/CL-PR-10-96-3.html

When you have finished reviewing my changes, please set the checked parameter below to true or failed to let others know (documentation at {{Sourcecheck}}).

This message was posted before February 2018. After February 2018, "External links modified" talk page sections are no longer generated or monitored by InternetArchiveBot. No special action is required regarding these talk page notices, other than regular verification using the archive tool instructions below. Editors have permission to delete these "External links modified" talk page sections if they want to de-clutter talk pages, but see the RfC before doing mass systematic removals. This message is updated dynamically through the template {{source check}} (last update: 5 June 2024).

If you have discovered URLs which were erroneously considered dead by the bot, you can report them with this tool.
If you found an error with any archives or the URLs themselves, you can fix them with this tool.

Cheers.—InternetArchiveBot (Report bug) 04:03, 2 December 2016 (UTC)[reply]

[Fisher1995-1] Dennis Fisher and I. J. Bahl (1995). Gallium Arsenide IC Applications Handbook. Vol. 1. Elsevier. p. 61. ISBN 978-0-12-257735-2. 'Clear search' to see pages

[1]