carlos@io.UUCP (Carlos Smith x4433) (02/04/89)
I have heard lots of wonderful things about Lattice's latest release. Manx has not had an update in a year, in spite of bugs (in fact, I just found a new one in the inline 68881 support). I need the maximum speed possible for the project I am working on. Are Lattice's claims of dramtically higher performance true? It would be worth it to switch if it is indeed 50% faster than Manx on compute intensive code. So, has anyone out there switched and ACTUALLY EXPERIENCED a dramatic improvement in the performance of their code? Do you find that compile time compares favorable with Manx? Does the floating point support seem robust? Anyone using the in-line floating point support? And, is the compiler stable, that is, pretty bug-free. I have heard the support is good, but I would prefer to not have to use it! If you did switch from Manx, were there any problems? What were the main incompatibilities? Also, I have heard very little about their debugger (CodeProbe?). Any comments on its interface? How well does it work with code that opens it's own screens, etc. I am also very interested in direct comparisons between it and SDB by anyone who has used both. Thanks for any information. As any professional programmer knows, productvity is everything, and that means using the best tools available. So, is Lattice now enough better to be worth the switch? -- Carlos Smith uucp:...!mit-eddie!ileaf!carlos Bix: carlosmith
jesup@cbmvax.UUCP (Randell Jesup) (02/07/89)
In article <937@io.UUCP> carlos (Carlos Smith x4433) writes: >I need the maximum speed possible for the project I am working on. Are >Lattice's claims of dramtically higher performance true? It would be worth >it to switch if it is indeed 50% faster than Manx on compute intensive code. The performance is much higher than before. I recompiled Moria, about 1 Meg of C, and cut the executable by > 20% in size. You can do better with other programs (Moria is generic, but often ugly, C translated from Pascal). >So, has anyone out there switched and ACTUALLY EXPERIENCED a dramatic >improvement in the performance of their code? Do you find that compile time >compares favorable with Manx? Does the floating point support seem robust? >Anyone using the in-line floating point support? And, is the compiler stable, >that is, pretty bug-free. I have heard the support is good, but I would prefer >to not have to use it! The 5.02 patches were just released, and are being sent to all registered owners. The improvement is fairly dramatic, far more so if you are using an '020/'030 or '881/'882. >Thanks for any information. As any professional programmer knows, productvity >is everything, and that means using the best tools available. So, is Lattice >now enough better to be worth the switch? In my mind, yes. Disclaimer: I beta-tested for Lattice, and have rarely used Manx. -- Randell Jesup, Commodore Engineering {uunet|rutgers|allegra}!cbmvax!jesup
papa@pollux.usc.edu (Marco Papa) (02/07/89)
In article <5909@cbmvax.UUCP> jesup@cbmvax.UUCP (Randell Jesup) writes: > The 5.02 patches were just released, and are being sent to all >registered owners. The improvement is fairly dramatic, far more so if you >are using an '020/'030 or '881/'882. Is this what LATTICE told you? Should we "registered" owners of Lattice C 5.0 wait for the upgrade in the mail, or should we log on the BBS and pick up the 5.02 patches? -- Marco Papa 'Doc' -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= uucp:...!pollux!papa BIX:papa ARPAnet:pollux!papa@oberon.usc.edu "There's Alpha, Beta, Gamma and Diga!" -- Leo Schwab [quoting Rick Unland] -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
cmcmanis%pepper@Sun.COM (Chuck McManis) (02/08/89)
In article <5909@cbmvax.UUCP> jesup@cbmvax.UUCP (Randell Jesup) writes: > The 5.02 patches were just released, and are being sent to all >registered owners. In article <15200@oberon.USC.EDU> papa@pollux.usc.edu (Marco Papa) responds: > Is this what LATTICE told you? Should we "registered" owners of Lattice > C 5.0 wait for the upgrade in the mail, or should we log on the BBS and > pick up the 5.02 patches? Yes, this is what LATTICE told us on BIX. Generally, I've got some mixed emotions on the whole 5.0 -> 5.02 deal here. On the one hand, they responded rather quickly to the bugs, on the other hand there were a few bugs that beta testing should have caught. Then they put the updates on BIX (this is a better deal for some people) but on the other hand they had essentially zip documentation to go with the 7 or 8 ARC files of fixes. Most use "Patch" (which is on the distribution disks) but patch isn't documented in either manual or mentioned in the Master Index. It has really been hot and cold emotions here. I spent three hours last night attempting to apply the fixes, doing so very carefully so that I would know that the correct modules had been fixed and that the fixed versions replaced the old versions. I'm still not convinced I got it all right. Yes, it is a wonderful compiler, and yes many of the fixes were appreciated, but gee, what a way to waste an evening. --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.
fnf@estinc.UUCP (Fred Fish) (02/08/89)
In article <88695@sun.uucp> cmcmanis@sun.UUCP (Chuck McManis) writes: > [stuff about patching Lattice 5.0 to get 5.02] > I spent >three hours last night attempting to apply the fixes, doing so very >carefully so that I would know that the correct modules had been fixed >and that the fixed versions replaced the old versions. I'm still not >convinced I got it all right. Yes, it is a wonderful compiler, and >yes many of the fixes were appreciated, but gee, what a way to waste >an evening. One simple way to ensure that their customers got all the patches right would be for them to include a crc listing of all the files on a patched disk. The "rls" and "crc" programs in my library solve exactly this problem. -Fred -- # Fred Fish, 1835 E. Belmont Drive, Tempe, AZ 85284, USA # 1-602-491-0048 asuvax!{nud,mcdphx}!estinc!fnf
mwm@eris.berkeley.edu (Mike (I'll think of something yet) Meyer) (02/11/89)
In article <59@estinc.UUCP> fnf@estinc.UUCP (Fred Fish) writes: <In article <88695@sun.uucp> cmcmanis@sun.UUCP (Chuck McManis) writes: <> [stuff about patching Lattice 5.0 to get 5.02] <> I spent <>three hours last night attempting to apply the fixes, doing so very <>carefully so that I would know that the correct modules had been fixed <>and that the fixed versions replaced the old versions. That's odd - I spent about that long; most of it was downloading the patch files. Then again, I'd also been through the 5.01 patches, so the documentation wasn't that important. Did you use the "official" update sent in the mail? Or did you download the patches? I wouldn't be surprised if the download patches are missing the documentation, on the expectation that they go to an "experienced" users. That doesn't excuse them from being missing, though. <One simple way to ensure that their customers got all the patches <right would be for them to include a crc listing of all the files <on a patched disk. The "rls" and "crc" programs in my library <solve exactly this problem. The lattice patch program checksums the result, and lets you know if you got it right. You then have to rename the patched.new file to patched (after renaming patched to patched.old, one hopes). That the libraries are patched in a different way, and some commmands were replaced instead of patched doesn't help. Since you have to have assigns pointing to the libraries, include files and binaries, they ought to be able to distribute a cli script that would patch everything for you. That should take care of all the above problems. BTW, LMK 1.04 (the 5.10 lmk) disagrees badly with LMK 1.05 (the 5.02) lmk. They broke LMK badly enough that I'm going back to make. <mike -- Il brilgue: les toves lubricilleux Mike Meyer Se gyrent en vrillant dans le guave, mwm@berkeley.edu Enmimes sont les gougebosqueux, ucbvax!mwm Et le momerade horsgrave. mwm@ucbjade.BITNET