home *** CD-ROM | disk | FTP | other *** search
- Xref: sparky comp.os.os2.programmer:4503 comp.os.os2.misc:28473
- Path: sparky!uunet!mcsun!uknet!cam-orl!root
- From: thg@cam-orl.co.uk (Tim Glauert)
- Newsgroups: comp.os.os2.programmer,comp.os.os2.misc
- Subject: Re: Does math coprocessor speed up GUI (here: OS/2 presentation mgr)?
- Message-ID: <1992Aug28.141752.23615@cam-orl.co.uk>
- Date: 28 Aug 92 14:17:52 GMT
- References: <17fuahINN1tl@iraul1.ira.uka.de> <19920826.074333.779@almaden.ibm.com> <19920827.085206.286@almaden.ibm.com>
- Sender: root@cam-orl.co.uk (Operator)
- Organization: Olivetti Research Ltd, Cambridge, England.
- Lines: 11
-
- In article <19920827.085206.286@almaden.ibm.com>, ADUNSMUI@TOROLAB6.VNET.IBM.COM (Al Dunsmuir) writes:
- |> Tim,
- |>
- |> My understanding is that PMX performs not only the standard MIT X server
- |> functions, but must translate the X-windows actions into the equivalent PM
- |> actions. Apparently this intermediate layer has the floating-point code in
- |> question.
-
- But this thread started with the claim that PM was fast because it was all
- integer. So these guys are coverting from int to float and then back to int
- again? Bizarre!
-