Content deleted Content added
Peter Flass (talk | contribs) m →External links: typo |
Peter Flass (talk | contribs) →Microcode: change heading level |
||
Line 66:
In response to competitive pressures, IBM introduced a memory upgrade option, allowing 96K on a 360/30.<ref>Historical Note: A posted submission to Ed Thelen's 360/30 writings (http://ed-thelen.org/comp-hist/ibm-360-30.html) signed LEONARD.BEARES.ctr (at) disa.mil (December 2005) indicated that, in response to a 128K offering by a small company, IBM actually introduced an upgrade, allowing 96K on a 360/30. BEARES noted that this allowed hour-plus tape/disk-based sorts to be done in-core in minutes.</ref> It seems, based on the system's front panel, that a provision for supporting more than 64K had been pre-planned.<ref>http://www.ljw.me.uk/ibm360/links has an entry named '''Glenn's Computer Museum''' (http://www.glennsmuseum.com/ibm/ibm.html) that has the following annotation: note that this front panel has indicators for extra address bits to allow more than 64k main storage</ref>
The [[CPU]] used an 8-bit [[microarchitecture]] with only a few [[hardware register]]s; everything that the programmer saw was
[[Emulator|emulated]] by the [[microprogram]].<ref name="30theory">{{cite book|url=http://bitsavers.org/pdf/ibm/360/fe/2030/Y24-3360-1_2030_FE_Theory_Opns_Jun67.pdf|title=Field Engineering Theory of Operation, 2030 Processing Unit, System/360 Model 30|id=Y24-3360-1|edition=Fifth|date=June 1967|publisher=IBM}}</ref><ref>{{cite book|url=http://bitsavers.org/pdf/ibm/360/fe/2030/Model_30_Microprogramming_Lang.pdf|title=Model 30 Microprogramming Language|publisher=IBM}}</ref> Handling a 4-byte word took (at least) 6 microseconds, based on a 1.5 microsecond storage access cycle time.<ref name="30theory"/>{{rp|pg.1–4}}
|