rwa@cs.AthabascaU.CA (Ross Alexander) (11/22/89)
For what it's worth: I have a blitter. It's $75 down the toilet. The speedup it provides is _very_ marginal ( < 25% ), and affects only a few operations (essentailly, large blits). QuikST and the TOS 1.4 upgrade are much more cost effective. When I get home tonight I'll try to post some benchmark numbers. Also for what it's worth: the JRI Jato board is another $100 down the toilet. (This is a soi-disant cpu-"speedup" board). I measure about 5 to 10% speedup. This is not enough to notice; it's measurable, but not _noticeable_. I agree a benchmark can be cooked up that this board will show much better results from, but in general use (which is what I generally do, of course) it's a looser. On a more upbeat note, I am encouraged by what I've seen so far of Dave Small's '030 board proposal, and hope he can make it into a product. Between this, the TT/p, and the TT/x, there may indeed be an upgrade path for ST users who need more horespower and better video after all. It sure has been a long time coming, though. 1985 --> 1990 is five years; in this business, that's about a half century. Ross
pa1329@sdcc13.ucsd.edu (pa1329) (11/23/89)
In article <1261@atha.AthabascaU.CA> rwa@cs.AthabascaU.CA (Ross Alexander) writes: >For what it's worth: I have a blitter. It's $75 down the toilet. >The speedup it provides is _very_ marginal ( < 25% ), and affects >only a few operations (essentailly, large blits). QuikST and the >TOS 1.4 upgrade are much more cost effective. When I get home >tonight I'll try to post some benchmark numbers. > Ross Hmmm,. I wonder why is the Atari blitter so ineffective. Its speed is no where comparable to the Amiga's blitter. Why is the difference?
daveh@cbmvax.UUCP (Dave Haynie) (11/23/89)
in article <5312@sdcc6.ucsd.edu>, pa1329@sdcc13.ucsd.edu (pa1329) says: > Hmmm,. I wonder why is the Atari blitter so ineffective. Its > speed is no where comparable to the Amiga's blitter. Why is the > difference? I know of two differences. The primary one is that the Atari blitter shares the main system bus with the 68000 in the ST. So you have to alternate between blitter access and CPU access. The Amiga blitter, on the other hand, is on a separate bus, and when the video load isn't outrageous, or the CPU has work to do outside of video memory, the CPU and blitter can operate in parallel. And since the Amiga blitter only takes two clocks for a memory cycle, it can run twice the speed of the 68000 if it uses both halves of the cycle (thus kicking out the 68000, but of course, the Atari always has to do that). The other obvious difference is that the Atari blitter handles only two sources into one destination, with 16 possible operations, while the Amiga blitter handler three sources into one destination, with 256 possible operations. There are some clever additions to plain bit blitting. The Atari blitter apparently has a little internal chunk of memory which is used to support pattern fills without having to constantly fetch the pattern from main memory. The Amiga blitter has hardware line drawing. -- Dave Haynie Commodore-Amiga (Systems Engineering) "The Crew That Never Rests" {uunet|pyramid|rutgers}!cbmvax!daveh PLINK: hazy BIX: hazy Too much of everything is just enough
mfolivo@sactoh0.UUCP (Mark F. Newton) (11/24/89)
In article <5312@sdcc6.ucsd.edu>, pa1329@sdcc13.ucsd.edu (pa1329) writes: > In article <1261@atha.AthabascaU.CA> rwa@cs.AthabascaU.CA (Ross Alexander) writes: > >For what it's worth: I have a blitter. It's $75 down the toilet. > >The speedup it provides is _very_ marginal ( < 25% ), and affects > >only a few operations (essentailly, large blits). QuikST and the > >TOS 1.4 upgrade are much more cost effective. When I get home > >tonight I'll try to post some benchmark numbers. > > Ross > > > > > Hmmm,. I wonder why is the Atari blitter so ineffective. Its > speed is no where comparable to the Amiga's blitter. Why is the > difference? Well, what in the world is he using it for? What do you want? I have a Mega4 and a 1040ST and I like the Blitter. "Compared to the Amiga blitter", please explain. Do you have both set up next to each other and running the same program? -Shinobu Sacramento, CA pacbell!sactoh0!mfolivo mmsac!sactoh0!mfolivo "Issho kakete iwasete miseru-cha" "Imawano kiwani itteyaru."
kbad@atari.UUCP (Ken Badertscher) (11/28/89)
pa1329@sdcc13.ucsd.edu (pa1329) writes: | Hmmm,. I wonder why is the Atari blitter so ineffective. "Ineffective" is a pretty strong word. As has been mentioned, software speeder-uppers like TurboST and QuickST give a dramatic speed improvement because they solve the real bottleneck: software overhead. The BLiTTER chip does what it was designed to do very well. That is, BitBlt operations are sped up tremendously. A large part of the graphics processing being done by many ST's is TEXT bit block transfer operations. As has been demonstrated by Wayne Buckholt and Darek Mihocka, it is possible, given room, to speed these up significantly. Fast, resolution specific blit dispatching and code combined with the BLiTTER chip is phenomenal. As an aside, I should point out that our VDI guy has put a lot of work into the blit code for the TT, and it's pretty phenomenal _without_ hardware assist. One other thing that will give a noticeable improvement is the lack of Line F compression in STE and TT ROMs. The Line F compression adds a bit of overhead to AES operations, and desktop/window/ dialog operations are noticeably sped up when the compression is gone. | Its speed is no where comparable to the Amiga's blitter. As I said above, I think the real problem is a software bottleneck. If it were possible to put the Ami blitter in ST's, you'd probably get similar results - albeit with a chip more powerful in terms of variety of operations. -- ||| Ken Badertscher (ames!atari!kbad) ||| Atari R&D System Software Engine / | \ #include <disclaimer>
rehrauer@apollo.HP.COM (Steve Rehrauer) (11/30/89)
In article <1819@atari.UUCP> kbad@atari.UUCP (Ken Badertscher) writes: >As an aside, I should point out that our VDI guy has put a lot of work >into the blit code for the TT, and it's pretty phenomenal _without_ >hardware assist. One other thing that will give a noticeable improvement >is the lack of Line F compression in STE and TT ROMs. The Line F >compression adds a bit of overhead to AES operations, and desktop/window/ >dialog operations are noticeably sped up when the compression is gone. "Duuuhhhh" (he says in embarassed stupefaction), "Whut's 'Line F compression'?" I know what you mean by "Line F", but what's the "compression" refer to re: ST ROMs? (Usually I'm only this dense on Mondays & national holidays...) -- >>"Aaiiyeeee! Death from above!"<< | Steve Rehrauer, rehrauer@apollo.hp.com "Flee, lest we be trod upon!" | The Apollo System Division of H.P.