Editing Cray-3 (section)

== History ==

===Background===
Seymour Cray began the design of the Cray-3 in 1985, as soon as the [[Cray-2]] reached production.{{sfn|Trew|2012|p=245}} Cray generally set himself the goal of producing new machines with ten times the performance of the previous models. Although the machines did not always meet this goal, this was a useful technique in defining the project and clarifying what sort of process improvements would be needed to meet it.{{sfn|MacKenzie|1998|p=141}} For the Cray-3, he decided to set an even higher performance improvement goal, an increase of 12x over the Cray-2.{{sfn|MacKenzie|1998|p=153}}

Cray had always attacked the problem of increased speed with three simultaneous advances; more [[execution unit]]s to give the system higher [[Parallel computing|parallelism]], tighter packaging to decrease signal delays, and faster components to allow for a higher clock speed. Of the three, Cray was normally least aggressive on the last; his designs tended to use components that were already in widespread use, as opposed to leading-edge designs.{{sfn|MacKenzie|1998|p=141}}

For the Cray-2, he introduced a novel 3D-packaging system for its [[integrated circuit]]s to allow higher densities,{{sfn|Readings|2000|p=10}} and it appeared that there was some room for improvement in this process. For the new design, he stated that all wires would be limited to a maximum length of {{convert|1|foot}}. This would demand the processor be able to fit into a {{convert|1|cuft}} block, about {{frac|3}} that of the Cray-2 CPU. This would not only increase performance but make the system 27 times smaller.{{sfn|Trew|2012|p=246}}

For a 12x performance increase, the packaging alone would not be enough, the circuits on the chips themselves would also have to speed up. The Cray-2 appeared to be pushing the limits of the speed of [[silicon]]-based [[transistor]]s at 4.1&nbsp;ns (244&nbsp;MHz), and it did not appear that anything more than another 2x would be possible. If the goal of 12x was to be met, more radical changes would be needed, and a "high tech" approach would have to be used.{{sfn|MacKenzie|1998|pp=153–154}}

Cray had intended to use [[gallium arsenide]] circuitry in the Cray-2, which would not only offer much higher switching speeds but also used less energy and thus ran cooler as well. At the time the Cray-2 was being designed, the state of GaAs manufacturing simply was not up to the task of supplying a supercomputer.{{sfn|Readings|2000|p=9}} By the mid-1980s, things had changed and Cray decided it was the only way forward.{{sfn|MacKenzie|1998|p=154}} Given a lack of investment on the part of large chip makers, Cray decided to invest in a GaAs chipmaking startup, GigaBit Logic, and use them as an internal supplier.<ref>{{cite news |first=James |last=Peltz |url=https://www.latimes.com/archives/la-xpm-1990-01-23-fi-494-story.html |title=GigaBit Logic Negotiating Sale With Cray Computers |newspaper=Los Angeles Times |date=23 January 1990}}</ref>

Describing the system in November 1988, Cray stated that the 12 times performance increase would be made up of a three times increase due to GaAs circuits, and four times due to the use of more processors. One of the problems with the Cray-2 had been poor multiprocessing performance due to limited [[Bandwidth (computing)|bandwidth]] between the processors, and to address this the Cray-3 would adopt the much faster architecture used in the [[Cray Y-MP]]. This would provide a design performance of 8000 [[Million instructions per second|MIPS]], or 16 [[FLOPS|GFLOPS]].{{sfn|Trew|2012|p=246}}

===Development===
The Cray-3 was originally slated for delivery in 1991.<ref name="CRAY-COMPUTER-CORPORATION-Mar-1996-8-K">{{cite web|url=http://edgar.secdatabase.com/2189/92735696000121/filing-main.htm |title=CRAY COMPUTER CORPORATION, Form 8-K, Current Report, Filing Date Mar 26, 1996 |publisher=secdatabase.com |access-date =May 14, 2018}}</ref> This was during a time when the supercomputer market was rapidly shrinking from 50% annual growth in 1980, to 10% in 1988.{{sfn|MacKenzie|1998|p=154}} At the same time, Cray Research was also working on the Y-MP, a faster multi-processor version of the system architecture tracing its ancestry to the original [[Cray-1]]. In order to focus the Y-MP and Cray-3 groups, and with Cray's personal support,{{sfn|Murray|1997|p=195}} the Cray-3 project moved to a new research center in [[Colorado Springs, Colorado|Colorado Springs]].{{sfn|Trew|2012|p=245}}

By 1989, the Y-MP was starting deliveries, and the main CRI lab in [[Chippewa Falls, Wisconsin]], moved on to the C90, a further improvement in the Y-MP series.{{sfn|Trew|2012|p=243}}{{sfn|MacKenzie|1998|pp=154–155}} With only 25 Cray-2s sold, management decided that the Cray-3 should be put on "low priority" development. In November 1988, the Colorado Springs lab was spun off as [[Cray Computer|Cray Computer Corporation]] (CCC), with CRI retaining 10% of the new company's stock and providing an $85 million promissory note to fund development.{{sfn|Trew|2012|p=245}} Cray himself was not a shareholder in the new company, and worked under contract.<ref name="www.nytimes.com business-people-chief-executive-quits-at-cray-computer">{{cite news |url=https://www.nytimes.com/1992/04/17/business/business-people-chief-executive-quits-at-cray-computer.html?src=pm |title=Chief Executive Quits At Cray Computer |newspaper=The New York Times |date=17 April 1992}}</ref>{{sfn|Murray|1997|p=190}} As CRI retained the lease on the original building, the new company had to move once again, introducing further delays.{{sfn|Trew|2012|p=245}}{{sfn|Readings|2000|p=10}}

By 1991, development was behind schedule.<ref>{{cite news |url=https://www.nytimes.com/1991/12/17/business/company-news-cray-computer-is-behind-schedule.html?src=pm |title=Cray Computer Is Behind Schedule |newspaper=The New York Times |date=17 December 1991}}</ref> Development slowed even more when [[Lawrence Livermore National Laboratory]] cancelled its order for the first machine,<ref>{{cite news |title=Cray Loses Only Order For Product |work=The New York Times |date=24 December 1991 |url=https://www.nytimes.com/1991/12/24/business/company-news-cray-loses-only-order-for-product.html?src=pm }}</ref> in favor of the C90. Several executives, including the CEO, left the company.<ref name="www.nytimes.com business-people-chief-executive-quits-at-cray-computer"/> The company then announced they would be looking for a customer that needed a smaller version of the machine, with four to eight processors.<ref name=cold/>

The first (and only) production model (serial number S5, named ''Graywolf'') was loaned to [[NCAR]] as a demonstration system in May 1993. NCAR's version was configured with 4 processors and a 128&nbsp;MWord (64-bit words, 1&nbsp;GB) common memory.{{sfn|Lester|1993}} In service, the [[static RAM]] proved to be problematic. It was also discovered that the [[square root]] code contained a bug that resulted in 1 in 60 million calculations being wrong. Additionally, one of the four CPUs was not running reliably.<ref>{{cite magazine |first=Gary |last=Anthes |title=Research lab sizes up slew of supercomputers |magazine=Computerworld |date=1 August 1994 |page=55 |url=https://books.google.com/books?id=mEU4Qex5594C&pg=PA55}}</ref>

CCC declared bankruptcy in March 1995, after spending about $300 million of financing. NCAR's machine was officially decommissioned the next day.<ref>{{cite web |title=CRAY-3 (graywolf): 1993–1995 |website=SCD Supercomputer Gallery |url=https://www.cisl.ucar.edu/computers/gallery/cray/graywolf.jsp |archive-url=https://web.archive.org/web/20161010004607/https://www.cisl.ucar.edu/computers/gallery/cray/graywolf.jsp |archive-date=2016-10-10 |access-date=2016-10-20 |url-status=dead }}</ref> Seven system cabinets, or "tanks", serial numbers S1 to S7, were built for Cray-3 machines. Most were for smaller two-CPU machines. Three of the smaller tanks were used on the [[Cray-4]] project,<ref>{{cite magazine |title=Seymore R. Cray |magazine=Computerworld |date=18 July 1994 |page=20 |url=https://books.google.com/books?id=atdXb3xr3vgC&pg=PA20}}</ref> essentially a Cray-3 with 64 faster CPUs running at 1&nbsp;ns (1&nbsp;GHz) and packed into an even smaller space.<ref name= Stedman>{{cite magazine |first=Craig |last=Stedman |title=Cray Computer ends its quest |magazine=Computerworld |date=3 April 1995 |page=32 |url=https://books.google.com/books?id=J-JEQiC7c3oC&pg=PA32}}</ref> Another was used for the [[Cray-3/SSS]] project.<ref>{{cite book |first= Reinhard |last=Wobst |title= Cryptology Unlocked |publisher= John Wiley & Sons |date= 2007 |page=[https://archive.org/details/Cryptology_Unlocked/page/n166 150] |isbn=9780470060643 |url=https://archive.org/details/Cryptology_Unlocked }}</ref>

The failure of the Cray-3 was in large part due to the changing political and technical climate. The machine was being designed during the collapse of the [[Warsaw Pact]] and ending of the [[Cold War]], which led to a massive downsizing in supercomputer purchases.<ref name=cold>{{cite news |url=https://www.nytimes.com/1992/02/21/business/cold-war-s-end-hits-cray-computer.html?src=pm |title=Cold War's End Hits Cray Computer |newspaper=The New York Times |date=21 February 1992}}</ref><ref name=wsj/> At the same time, the market was increasingly investing in [[massively parallel]] (MP or MPP) designs. Cray was critical of this approach, and was quoted by ''[[The Wall Street Journal]]'' as saying that MPP systems had not yet proven their supremacy over vector computers, noting the difficulty many users have had programming for large parallel machines. "I don't think they'll ever be universally successful, at least not in my lifetime".<ref name=wsj>{{cite news |first=Michael |last=Allen |title=Pushing Big Iron: Seymour Cray's Woes Reflect Tough Times for Supercomputers |newspaper=Wall Street Journal |date= 1998 |url=http://www.isip.piconepress.com/publications/courses/msstate/ece_8463/projects/1998_spring/data/lm_training/wsj93_042.text |archive-url=https://web.archive.org/web/20120317081044/http://www.isip.piconepress.com/publications/courses/msstate/ece_8463/projects/1998_spring/data/lm_training/wsj93_042.text |archive-date=17 March 2012}}</ref>