nw@palladium.UUCP (Neil Webber) (07/20/87)
I can't recall having ever seen a posting of Fish Disk (Rick Spanbauer) benchmark results for the 40MB Pal Jr. disk, so here they are. Note that the 40MB drive has a faster seek time than the 20MB drive, hence the numbers come out a little different: File creations: 10/sec File deletions: 28/sec Directory scan: 41/sec Seek/read: 79/sec Read 512 byte file: 40960 bytes/sec Write 512 byte file: 15065 bytes/sec Read 4096 byte file: 59578 bytes/sec Write 4096 byte file: 22598 bytes/sec Read 8192 byte file: 59578 bytes/sec Write 8192 byte file: 23616 bytes/sec Read 32768 byte file: 60963 bytes/sec Write 32768 byte file: 24499 bytes/sec Having a hard disk and some real memory (3.5MB) has made my Amiga a new machine! (It has also "initialized" my bank account). Now, if I could just teach Soundscape to not crash just because it wasn't started from its copy-protected floppy. What was the number of that pirate bulletin board? :-) :-) :-) :-) -- Neil Webber / Palladium Data Systems / (No longer confused with Jordan Marsh) gilaabpdywyc! / Marlboro MA, 01752 {linus!alliant, harvard!cfisun}!palladium!nw
kent@xanth.UUCP (Kent Paul Dolan) (07/22/87)
In article <337@palladium.UUCP> nw@palladium.UUCP (Neil Webber) writes: >[..] Fish Disk (Rick Spanbauer) benchmark results for the 40MB Pal Jr. disk >[...] > Read 32768 byte file: 60963 bytes/sec > Write 32768 byte file: 24499 bytes/sec No complaint against the Pal Jr., but just generically; with the state of the art in hard disk file transfer in the multiple megabytes per second range, why are we still down in the 10K's of bytes? Speeds this low make disk based virtual memory a real loser (although at today's prices, extended address space ramdisk looks like a real possibility for virtual memory on PC's with limited personal memory, like the Amiga). To put it another way, where are the bottlenecks now? Kent, the man from xanth.
doc@crash.CTS.COM (Mitch Evans) (07/22/87)
Howdy Foolks! I am interested in finding out if the Amiga series is capable of becoming a Net (UUCP, ARPA, or otherwise) site. Is there any hardware needed? If there a package out there somewhere? What else should I know about it? Additionally, if there is no package out there, I am VERY willing/interested in the writing of one. If anyone in the San Diego area would like to be in on this, leave me net mail, or call me at (619) 588-9247/ask for Mitch. Doc -- ******************************************************************************* UUCP: {cbosgd,hplabs!hp-sdd,sdcsvax,nosc}!crash!doc ARPA: crash!doc@nosc.mil INET: doc@crash.CTS.COM .........Quid?.....Me Vexari?.. ********************************************************************************
cmcmanis%pepper@Sun.COM (Chuck McManis) (07/23/87)
In article <1638@xanth.UUCP> kent@xanth.UUCP (Kent Paul Dolan) writes: >No complaint against the Pal Jr., but just generically; with the state of the >art in hard disk file transfer in the multiple megabytes per second range, >why are we still down in the 10K's of bytes? Ok, the spec for *any* ST-506 compliant drive is 5 megaBITS per second which is 625K bytes per second. For ESDI drives the data transfer rate is 10 megaBITS per second or 1.25 MegaBYTES per second. Disk to host adapters (like a lot of SCSI thingies are) are usually capable of 1.2 megabyte capable with speeds up to 4 megabytes/second when run 'synchronously'. They basically buffer up the requested data in an on-board buffer and then squirt it out to you when they have it all. Some of the new SCSI chips can go faster than this. The Amiga's fastest transfer rate for bytes is about 600K bytes per second ( mov.b (a0)+,(a1)+, jmp $ ) 1.2 meg if you transfer words. Disks with DMA channels win big in this situation, although only 2X to 3X in raw transfer rate so that is the absolute best you could hope for. Now tack on the over -head of going from a request for a block in the file system and which is translated into a device request which is translated into a physical request which is copies back into the address space of the user and then notified. It can be done faster, but don't ever expect 'multiple megabytes' per second on a stock Amiga. > Speeds this low make disk based >virtual memory a real loser (although at today's prices, extended address >space ramdisk looks like a real possibility for virtual memory on PC's with >limited personal memory, like the Amiga). This doesn't parse for me. If you put RAM on the Amiga you can use it, are you thinking about x86 type machines? --Chuck McManis uucp: {anywhere}!sun!cmcmanis BIX: cmcmanis ARPAnet: cmcmanis@sun.com These opinions are my own and no one elses, but you knew that didn't you.
fgd3@jc3b21.UUCP (Fabbian G. Dufoe) (07/24/87)
In article <1424@crash.CTS.COM>, doc@crash.UUCP writes: > > Howdy Foolks! > > I am interested in finding out if the Amiga series is capable of becoming a > Net (UUCP, ARPA, or otherwise) site. Is there any hardware needed? If there > a package out there somewhere? What else should I know about it? Given the amount of traffic on Usenet, you'll need a hard disk to handle it. Sounds like a nice idea, though. If you find or write the necessary software I'd like to know about it. --Fabbian Dufoe 350 Ling-A-Mor Terrace South St. Petersburg, Florida 33705 813-823-2350 UUCP: ...seismo!akgua!usfvax2!jc3b21!fgd3