TOPIC | 1941.0 | Mvax 40x faster than axp? |
| 1941.3 | Fri Dec 30 1994 18:01 | Smarter? Maybe DFVV is an idiot savant. | 46 lines |
| 1941.8 | Thu Jan 05 1995 20:04 | Sam Salesman has ben TFSOed | 32 lines |
TOPIC | 1980.0 | GOTO to end of DO-loop traps you in loop... |
| 1980.4 | Sun Jan 29 1995 17:45 | A problem that took 10 years to fix right | 32 lines |
TOPIC | 1982.0 | ACCVIO with DFAV V6.2.2 |
| 1982.2 | Mon Jan 30 1995 11:27 | See also 1980.* | 1 lines |
TOPIC | 1996.0 | HPARITH - integer overflow |
| 1996.6 | Tue Feb 07 1995 14:51 | 1 | 23 lines |
| 1996.8 | Tue Feb 07 1995 18:21 | Overflow effects compiler optimizations | 32 lines |
TOPIC | 2004.0 | SAVE differs on ALPHA from VAX |
| 2004.9 | Mon Feb 13 1995 12:45 | A Legacy IBM FORTRAN Feature | 15 lines |
| 2004.11 | Mon Feb 13 1995 14:12 | I remember differently | 16 lines |
| 2004.12 | Mon Feb 13 1995 14:28 | Digital documents "support" of saved array dummys | 23 lines |
TOPIC | 2074.0 | Optimize bug in DEC Fortran V3.7-680 |
| 2074.4 | Wed Mar 29 1995 11:44 | Not DAXPY but timing statistics doing the divide | 29 lines |
TOPIC | 2077.0 | 0 for continuation character |
| 2077.1 | Wed Mar 29 1995 18:25 | Zero is not a continuation character | 4 lines |
TOPIC | 2116.0 | yet another "GEM ASSERTION, Compiler internal error" |
| 2116.17 | Tue Apr 25 1995 20:14 | | 11 lines |
TOPIC | 1922.0 | test_and set optimization and cache coherence |
| 1922.1 | Mon Aug 07 1995 11:07 | Do not loop with ldl_l | 28 lines |
TOPIC | 1954.0 | What is a sst instruction? |
| 1954.2 | Fri Aug 11 1995 10:55 | Not sts but *stt* | 2 lines |
TOPIC | 2004.0 | Alpha Assembler Question |
| 2004.1 | Sat Aug 12 1995 22:50 | Most rules have exceptions | 31 lines |
TOPIC | 2313.0 | about HUGE_VAL in math.h |
| 2313.2 | Wed Sep 06 1995 11:37 | Needing "with_inexact" requires program rewriting | 22 lines |
TOPIC | 3473.0 | linpack |
| 3473.3 | Mon Nov 27 1995 16:59 | Don't forget -tune ev5 in addition to -O5 | 5 lines |
TOPIC | 3942.0 | which function called this function |
| 3942.2 | Wed Jan 10 1996 18:59 | Try the documentation set | 2 lines |
TOPIC | 4661.0 | Listing Archive Contents for Shared Object Files |
| 4661.2 | Fri Mar 08 1996 10:48 | Never use -ieee_with_inexact | 21 lines |
TOPIC | 7532.0 | Mon Oct 21 1996 15:40 | How do I turnoff DECnet mail Mime encoding | 16 lines |
| 7532.2 | Tue Oct 22 1996 11:45 | Whereis DNU_OSI? | 1 lines |
TOPIC | 7798.0 | -ieee and -float problems DUv4.0 |
| 7798.1 | Mon Nov 11 1996 10:26 | Source code please | 1 lines |
TOPIC | 8348.0 | VAX FP Data Type, VMS time format? |
| 8348.1 | Fri Jan 03 1997 18:31 | I hope your customer uses Fortran. | 12 lines |
TOPIC | 8841.0 | Continuing execution after SEGV? |
| 8841.2 | Sat Feb 15 1997 17:46 | sc_pc in sig_context structure | 14 lines |
TOPIC | 2519.0 | double precision questions ??? |
| 2519.15 | Wed May 17 1995 19:27 | Some technical nits | 18 lines |
TOPIC | 2565.0 | rounding problem - Hparith ?? |
| 2565.6 | Wed Dec 20 1995 15:11 | Fix the bug! Don't paper over it. | 23 lines |
| 2565.8 | Mon Jan 15 1996 11:43 | See the Alpha AXP Architecture Reference Manual | 21 lines |
TOPIC | 2599.0 | /OPT=LEVEL=5 |
| 2599.3 | Sun Mar 24 1996 20:29 | Pascal documentation should be different from Fortran documentation | 22 lines |
TOPIC | 2603.0 | value accepted in debugger but not runtime |
| 2603.9 | Wed Apr 17 1996 23:10 | Run-time versus compile-time rounding problems | 20 lines |
| 2603.10 | Wed Apr 17 1996 23:28 | Someone is rounding the wrong direction | 6 lines |
TOPIC | 2627.0 | Evaluation Order Dependendcies |
| 2627.18 | Fri Nov 08 1996 13:44 | Customer is always right | 58 lines |
TOPIC | 2630.0 | Obscure restriction on base types of pointers, generates confusing error message: UNDECLID |
| 2630.3 | Wed Oct 16 1996 12:34 | A more complicated example | 30 lines |
TOPIC | 2648.0 | Pascal Compile Problem - URGRNT ... |
| 2648.2 | Fri Dec 06 1996 17:00 | | 2 lines |
TOPIC | 2651.0 | New Year's Wish: preprocessor, more flexible OPEN |
| 2651.3 | Wed Jan 15 1997 15:13 | Comments on #1 and #3 | 13 lines |
TOPIC | 2670.0 | Multple PSECT contributions in same module - expected layout? |
| 2670.2 | Fri May 02 1997 11:19 | Some front ends lock down psect allocation | 11 lines |
TOPIC | 288.0 | Price Reductions on Alpha Microprocessors |
| 288.33 | Sun Oct 01 1995 17:03 | Linux should be ABI compatible with Digital Unix | 17 lines |
TOPIC | 354.0 | South Queensferry sold ? |
| 354.7 | Wed Jan 04 1995 18:26 | Receiving cash now is better than spending it later | 22 lines |
TOPIC | 420.0 | out-of-order execution, why not? |
| 420.8 | Thu Aug 03 1995 11:20 | More on compiler speculation | 50 lines |
TOPIC | 429.0 | Do Alpha chips ever produce non-finites? |
| 429.1 | Wed Sep 06 1995 22:04 | Not usually | 22 lines |
| 429.3 | Mon Sep 11 1995 12:25 | No implicit TRAPB between FP compute and store | 10 lines |
TOPIC | 450.0 | What makes a P6 fast? |
| 450.22 | Tue Jan 09 1996 20:40 | Motorola not Intel found error | 25 lines |
TOPIC | 476.0 | 21164/350 - EV56 - FAB6 / 0.35�m ??? |
| 476.19 | Mon Feb 12 1996 11:56 | The world looked different in 1988 | 49 lines |
TOPIC | 571.0 | HP PA8000 |
| 571.77 | Wed Oct 30 1996 14:31 | | 62 lines |
TOPIC | 596.0 | Alpha PC at COMDEX |
| 596.49 | Thu Dec 05 1996 09:42 | Does Digital *buy* D.UNIX or Alpha NT on desktops? | 8 lines |
TOPIC | 623.0 | ev6 & "out of order execution" feature |
| 623.10 | Mon Jan 06 1997 14:46 | Part# EC-Y7279-72 | 3 lines |
TOPIC | 630.0 | SPECint for Alpha? |
| 630.32 | Wed Jan 15 1997 10:32 | Old vs. New software | 37 lines |
TOPIC | 680.0 | Java and Alpha |
| 680.4 | Tue Apr 15 1997 12:40 | Two definitions of "independent" | 23 lines |
TOPIC | 688.0 | Where can I find MVI docs? |
| 688.1 | Mon May 05 1997 20:05 | AAH published last Fall | 2 lines |
TOPIC | 1065.0 | URGENT!!!!!!!$$$$$$ C RELARTED QUESTION FOR TES |
| 1065.1 | Tue Jan 17 1995 20:41 | Do man uswitch | 8 lines |
TOPIC | 1109.0 | Possible bug in DECC T5.0-000 |
| 1109.3 | Fri Feb 10 1995 18:35 | Addresses are signed in the AXP architecture | 13 lines |
TOPIC | 1114.0 | OSF: Intrinsics for decc on osf |
| 1114.3 | Tue Feb 14 1995 10:17 | Need __ADD_ATOMIC which returns updated value | 14 lines |
TOPIC | 1166.0 | Fortran calling C - changing descriptor addresses |
| 1166.6 | Thu Mar 16 1995 21:00 | Tartan Labs knows how to do this | 29 lines |
| 1166.11 | Fri Apr 07 1995 18:05 | Been there. Done that. | 30 lines |
TOPIC | 1175.0 | Question on behavior on ftn file |
| 1175.4 | Mon Mar 20 1995 16:09 | A Unix system would see two lines, not one line | 34 lines |
| 1175.6 | Wed Mar 22 1995 10:35 | Unix semantics are relevant for me | 23 lines |
TOPIC | 1240.0 | too many load/store |
| 1240.7 | Tue May 09 1995 17:55 | Infrequently called routine (caused by feedback) | 11 lines |
| 1240.8 | Tue May 09 1995 18:01 | Of course, the feedback could be correct!! | 5 lines |
TOPIC | 1478.0 | Effect of "-tune" switch... |
| 1478.1 | Thu Oct 05 1995 19:46 | Cross-posted in TURRIS::FORTRAN note 337 | 2 lines |
TOPIC | 1516.0 | SP changed, exception returns via SYS$GOTO_UNWIND, SP is restored wrong |
| 1516.7 | Tue Nov 07 1995 19:28 | SP should keep most recent setting after an unwind | 28 lines |
TOPIC | 1594.0 | Different results cc (ANSI) versus IBM,SGI... |
| 1594.8 | Fri Dec 15 1995 15:45 | Try -ieee_with_no_inexact | 3 lines |
TOPIC | 1653.0 | atof(), OSF/VMS behavior, compile-time constant vs. run-time behavior |
| 1653.1 | Sat Feb 03 1996 21:05 | Surprise! Both answers are correct. | 38 lines |
| 1653.2 | Sat Feb 03 1996 21:19 | Details from the analysis of .1 | 53 lines |
TOPIC | 1696.0 | Long double whith 128 bits size still wanted |
| 1696.9 | Fri Apr 26 1996 10:42 | Please don't have incompatible libraries | 43 lines |
TOPIC | 1768.0 | Need help with bit fields under DU. |
| 1768.1 | Fri Apr 19 1996 12:08 | Big-endian vs. Little-endian | 30 lines |
TOPIC | 1776.0 | Problem with sscanf |
| 1776.1 | Wed Apr 24 1996 15:57 | Source program bug--double/float mismatch | 22 lines |
TOPIC | 1781.0 | Question about EV5 performance |
| 1781.1 | Fri Apr 26 1996 17:15 | TUNE=EV5 works on EV4 hosted compilers | 19 lines |
TOPIC | 1956.0 | Write_rnd/ Read_rnd : How do they work? |
| 1956.3 | Wed Oct 02 1996 14:23 | Intitial rounding mode has changed over time | 12 lines |
TOPIC | 1988.0 | -xtaso_short -taso 2GB limit ? |
| 1988.6 | Tue Oct 29 1996 11:27 | Read the calling standard book | 57 lines |
TOPIC | 980.0 | Second TRAPB instruction removed |
| 980.2 | Thu Aug 17 1995 15:40 | Use EXCB instead of TRAPB | 11 lines |
TOPIC | 1023.0 | Wed Oct 11 1995 16:59 | DEC C RTL for Alpha VMS prints denormals incorrectly | 49 lines |
| 1023.2 | Wed Oct 18 1995 20:42 | Use VMS V6.1 | 2 lines |
TOPIC | 1058.0 | Incorrect, unwarned floating point conversion causes wrong float output. |
| 1058.4 | Fri Dec 15 1995 14:42 | QAR scanf and c89 | 21 lines |
TOPIC | 1060.0 | Fri Dec 15 1995 17:35 | -float should effect old-style float parameters | 74 lines |
| 1060.4 | Mon Dec 18 1995 16:50 | Half a loaf or a whole loaf???? | 24 lines |
| 1060.7 | Tue Dec 19 1995 11:22 | But are two loaves really better than one? | 13 lines |
TOPIC | 1069.0 | Thu Jan 11 1996 14:19 | No -O5 in listing switch dump | 3 lines |
TOPIC | 1210.0 | SIGFPE AND IEEE FLOAT |
| 1210.5 | Thu Oct 24 1996 20:04 | Full IEEE FP avoids trapping | 12 lines |
TOPIC | 1231.0 | Fri Nov 29 1996 17:32 | /IEEE=xxx does not change default from /FLOAT=G_FLOAT | 37 lines |
| 1231.7 | Wed Dec 04 1996 16:44 | Severity not as important as existence of warning message | 13 lines |
TOPIC | 1271.0 | Wed Mar 12 1997 18:19 | Internet bug report | 40 lines |
TOPIC | 108.0 | Help on optimizing |
| 108.4 | Fri Jun 16 1995 19:25 | -O5 requires active measurements | 15 lines |
TOPIC | 184.0 | Exception Handling |
| 184.14 | Fri Jul 21 1995 17:21 | IEEE divide by zero usually results in an Infinity | 10 lines |
| 184.17 | Sun Jul 23 1995 18:24 | May not work on future Alpha chips | 51 lines |
| 184.18 | Sun Jul 23 1995 18:46 | LIB$ESTABLISH is dynamic on both Alpha and VAX | 23 lines |
TOPIC | 192.0 | Sun Jul 23 1995 18:57 | Suggestion: no pointer alias directive | 10 lines |
TOPIC | 237.0 | Alcor vs SGI and DEC 3900 |
| 237.15 | Tue Aug 22 1995 11:14 | Rewrite copy loop | 45 lines |
| 237.16 | Tue Aug 22 1995 11:26 | Another poor loop | 20 lines |
TOPIC | 325.0 | Customer wants -fpe2 behaviour, at -fpe0 speeds |
| 325.14 | Tue Oct 03 1995 19:49 | We could have generated better code | 17 lines |
TOPIC | 592.0 | Sun Feb 04 1996 14:06 | Inconsistent Floating Point conversions | 58 lines |
| 592.3 | Mon Feb 05 1996 19:40 | The same but better has also been done | 23 lines |
TOPIC | 593.0 | dfvv v6.3-141 - possible optimizer bug? |
| 593.2 | Tue Feb 06 1996 14:05 | Nonstandard variable access probably need VOLATILE | 20 lines |
TOPIC | 625.0 | Prefetching? |
| 625.6 | Wed Feb 21 1996 14:24 | | 37 lines |
| 625.8 | Wed Feb 21 1996 18:49 | How prefetching helps | 30 lines |
| 625.14 | Fri Mar 29 1996 20:10 | asm cannot be used to prefetch | 7 lines |
| 625.16 | Fri Aug 23 1996 20:48 | More prefetching coming soon | 5 lines |
TOPIC | 693.0 | INTEGER *8 IN OPENVMS ALPHA FORTRAN |
| 693.7 | Wed Mar 27 1996 20:01 | 32-bit addresses are *sign* extended | 24 lines |
TOPIC | 709.0 | runtime library not initialized?? |
| 709.5 | Tue Apr 02 1996 22:53 | Unix has LIB$INITIALIZE | 11 lines |
TOPIC | 747.0 | Validating REAL data on input (e.g., checking for denormalized numbers) |
| 747.6 | Thu Apr 25 1996 15:50 | EXP does it right, why can't READ | 32 lines |
TOPIC | 797.0 | -fpe1 performance penalty |
| 797.14 | Sun Jun 02 1996 19:03 | Making implementation tradeoffs cannot make everyone happy | 49 lines |
TOPIC | 807.0 | Scalapack and denorm numbers |
| 807.13 | Mon Jun 03 1996 15:47 | Does we need an additional floating point mode? | 31 lines |
TOPIC | 843.0 | speculate |
| 843.4 | Tue Jun 25 1996 22:12 | The name __speculate_all is a phantom (to me) | 23 lines |
| 843.6 | Wed Jun 26 1996 12:50 | Possible ld/library version skew?? | 27 lines |
TOPIC | 872.0 | Fp exceptions, IEEE and performance |
| 872.14 | Wed Jul 17 1996 12:14 | Also, more than lifetimes | 18 lines |
TOPIC | 959.0 | Better compiled code?? |
| 959.2 | Wed Sep 04 1996 21:16 | We need an ISO standard for -O levels | 20 lines |
TOPIC | 986.0 | amod - another fp precision issue |
| 986.4 | Mon Sep 23 1996 19:21 | Digital's answer is exact--HP's is wrong in the last bit | 16 lines |
TOPIC | 1050.0 | Anomalous T7.1-75 performance - tune:ev5 and/or level:5 are losers |
| 1050.4 | Thu Nov 14 1996 16:50 | Working..(but generic no longer is ev4) | 19 lines |
TOPIC | 1188.0 | Why this accuracy problem ? |
| 1188.5 | Thu Feb 20 1997 11:12 | Does Sun's output have few digits? | 4 lines |
TOPIC | 1259.0 | fortran 2 decc..in alpha vms? |
| 1259.7 | Wed Apr 16 1997 19:44 | RTL environment might be an issue | 3 lines |
TOPIC | 1266.0 | Double and Signle precision performance issue |
| 1266.1 | Mon Apr 21 1997 11:33 | Probably memory bandwidth--not float precision | 16 lines |
TOPIC | 1277.0 | prefetching |
| 1277.5 | Wed Apr 30 1997 21:09 | Thanks for the .4 example | 17 lines |
TOPIC | 1304.0 | named common block problem on DFNT |
| 1304.8 | Thu May 22 1997 16:13 | COFF and ELF standards cause this problem | 28 lines |
TOPIC | 256.0 | Floating point formats (bit pattern)? |
| 256.9 | Sat May 06 1995 15:10 | VAX float assembly available on Unix | 15 lines |
TOPIC | 352.0 | /align=vax in compiler |
| 352.6 | Tue May 23 1995 21:37 | Ask the Pascal folks | 1 lines |
TOPIC | 490.0 | EV5 - EV4 Compiles |
| 490.6 | Fri Jul 07 1995 11:04 | Default on ev5 is still -tune ev4 | 16 lines |
TOPIC | 575.0 | Setting IEEE dynamic rounding bits in FPCR on OpenVMS |
| 575.4 | Wed Aug 16 1995 22:33 | fpcr and fp_c are different registers | 37 lines |
| 575.6 | Thu Aug 17 1995 15:19 | User mode can mainpulate fpcr | 7 lines |
TOPIC | 648.0 | Problem with IEEE floating point format |
| 648.3 | Sun Sep 17 1995 15:40 | IEEE Floating Point details are very tricky | 37 lines |
TOPIC | 808.0 | Memory barriers: MB, WMB, and EV5 |
| 808.3 | Wed Nov 29 1995 18:17 | WMB can be used for parallel, pipelined computations | 51 lines |
TOPIC | 91.0 | OSF Proposal: 2-bits in Code Range Descr. for Speculative Execution |
| 91.2 | Fri Apr 28 1995 15:38 | Need detailed proposal | 3 lines |
TOPIC | 92.0 | AVMS ENTRY_LENGTH revisited et al |
| 92.9 | Tue Feb 21 1995 19:13 | Let's keep future options open | 47 lines |
TOPIC | 95.0 | Is GCC stack probe sequence correct? |
| 95.1 | Tue Mar 14 1995 22:30 | GCC is not following the Calling Standard | 43 lines |
| 95.4 | Wed Mar 15 1995 21:57 | The calling standard could be clearer | 29 lines |
TOPIC | 97.0 | WNT Call Std ECO 5 re Segment Procedure Descriptor |
| 97.7 | Sun Apr 09 1995 15:38 | Yes | 1 lines |
TOPIC | 99.0 | Is this optimization allowed by the call standard? |
| 99.2 | Fri Apr 14 1995 21:23 | Which routine gets the exception? | 9 lines |
TOPIC | 100.0 | Unix Call Std ECO 4 re Speculative Execution |
| 100.3 | Mon May 15 1995 12:46 | Yes | 0 lines |
TOPIC | 103.0 | Tue Nov 07 1995 19:38 | VMS ECO 45 re changing SP via LIB$PUT_INVO_REGISTERS | 6 lines |
| 103.3 | Sun Nov 19 1995 17:32 | How does it do that??? | 15 lines |
| 103.6 | Tue Nov 28 1995 10:07 | SP ought not to be settable | 12 lines |
| 103.9 | Wed Nov 29 1995 18:35 | Delayed change of the SP is doable, but ... | 18 lines |
TOPIC | 105.0 | R28 is unpredictable after (external?) transfer |
| 105.2 | Fri Oct 04 1996 15:51 | What defines a procedure call/return?? | 12 lines |
| 105.10 | Tue Oct 08 1996 20:28 | | 5 lines |
TOPIC | 107.0 | Need description of fiber safety |
| 107.3 | Sun Nov 03 1996 16:33 | Floating point context | 9 lines |
TOPIC | 108.0 | Must all code be "covered" by some descriptor? |
| 108.3 | Thu Jan 30 1997 11:35 | Need a rule for legacy object and images | 11 lines |
| 108.15 | Thu Feb 13 1997 13:42 | Remember inline parameters on 7094 and pdp-10 | 18 lines |
TOPIC | 109.0 | What alignment is assumed for global/external variables? |
| 109.15 | Fri Feb 07 1997 11:29 | Need to update Assembler manual | 18 lines |
| 109.20 | Fri Feb 07 1997 20:02 | ECO should be "invariant" rule | 39 lines |
| 109.29 | Thu Feb 13 1997 13:28 | | 14 lines |
| 109.33 | Sat Feb 15 1997 17:00 | | 12 lines |
| 109.37 | Wed Feb 19 1997 11:42 | Post deadline nit picking | 26 lines |