Activity for Philipp Klaus Krause
-
Philipp Klaus Krause modified a comment on ticket #714
Updated list of instruction based on work in Rabbit code generation done since. For most of these instructions, we have now prepared support in r4k codegen, so we could start testing them easily. For the few in the list, where codegen isn't prepared yet, I included them since I consider it likely that we'll start using them before SDCC 4.6.0. and iy, de call (hl) call (iy) cbm #n clr hl cp hl, #d cp hl, de dwjnz label ex bc, hl flag cc, hl ; should this be flag cc, hl or flag hl, cc? ibox a jp cx,...
-
Disable test case that fails due to lackof optimizations (RFE #1013).
-
Moved to feature request, since this is due to running out of memory. So we need better optimizations to pass the test.
-
Ticket moved from /p/sdcc/bugs/3934/ Can't be converted: _category: other
-
hc008/s08 fails 64-bit rotation test
-
Yes, this is fixed now, and the ticket can be closed.
-
Enable use of further R4K instructions in r4k/r5k/r6k codegen.
-
Merge from trunk.
-
unsigned 16x16->32 multiplication support routine for r2ka and r3ka.
-
Prepare for 0.8 release
-
Enable use of further R4K instructions in r4k/r5k/r6k codegen.
-
Enable use of further R4K instructions in r4k/r5k/r6k codegen.
-
Enable use of further R4K instructions in r4k/r5k/r6k codegen.
-
Enable use of further R4K instructions in r4k/r5k/r6k codegen.
-
Enable use of further R4K instructions in r4k/r5k/r6k codegen.
-
Merges from trunk.
-
Enable use of further R4K instructions in r4k/r5k/r6k codegen.
-
Merge from trunk.
-
Enable use of further R4K instructions in r4k/r5k/r6k codegen.
-
Enable use of further R4K instructions in r4k/r5k/r6k codegen.
-
Enable use of further R4K instructions in r4k/r5k/r6k codegen.
-
sdasrab treats xor hl, de as if it was in .r4k_11 mode
-
Enable use of further R4K instructions in r4k/r5k/r6k codegen.
-
Enable use of further R4K instructions in r4k/r5k/r6k codegen.
-
neg hl in R4K uCsim missing
-
uCsim cp a, d(ix) disassembly
-
Save register pair jk in r4k/r5k/r6k interrupt handlers.
-
Use R4K 32-bit push/pop.
-
Z80 instruction timing
-
Enable use of a first Rabbit 4000 instruction (mulu) in r4k/r5k/r6k port.
-
Merges from trunk.
-
Preparations for use of Rabbit instruction mode 10 in r4k/r5k/r6k codegen.
-
Preparations for use of Rabbit instruction mode 10 in r4k/r5k/r6k codegen.
-
Merges from trunk.
-
Remove redundant emission of assembler directives.
-
Setup for use of Rabbit instruction mode 10.
-
Merge from trunk.
-
Revert a change that broke ds390.
-
Emit .r3ka instead of .r3k directive in Rabbit 3000A and later asm.
-
Merges from trunk.
-
Since this is already in the upstream source, I'd recommend to just use #pragma disable_warning in tests/serpent.c.in.
-
z80n and z180 failures in farm since pqc branch merge
-
Fixed in [r16176].
-
Merge pqc branch.
-
Merges from upstream branches.
-
Fix an issue in 16x16->32 multiplication support function.
-
Z180 uCsim not showing disassembly for mlt de
-
Fix a typo in an error message.
-
Fix an issue in code generation for shift / rotation.
-
Both the optimization (by having a 32x32->64 multiplication in the library) and a timeout increase by 50% are now in the pqc branch.
-
Both the optimization (by having a 32x32 multiplication in the library) and a timeout increase by 50% are now in the pqc branch.
-
Both the optimization and a timeout increase by 50% are now in the pqc branch.
-
Merges from trunk.
-
Increase timeouts to accomodate dilithium-ntt regression test.
-
Multiplication improvements inspired by kyber-ntt regression test.
-
The bottleneck is 32x32->64 multiplication. We could about double the speed of that one by having a dedicated library function for it. If we also double the timeout then, it should work.
-
The .ihx is the same. Looks like we just run into the simulator timeout for machines that do not have good single-thread performance. For z80/z80n, the test takes about 30s on a Ryzen AI Max 395+, but about 7 min on Power 9.
-
z80n and z180 failures in farm since pqc branch merge
-
Apparently, the same happen on ppc64 host.
-
Handle pre-C23 host compiler.
-
Disable test case that fails due to known bug #3928.
-
Ticket moved from /p/sdcc/feature-requests/1012/ Can't be converted: _milestone:
-
ds390 fails 64-bit rotation test
-
Updated list of instruction based on work in Rabbit code generation done since. For most of these instructions, we have now prepared support in r4k codegen, so we could start testing them easily. For the few in the list, where codegen isn't prepared yet, I included them since I consider it likely that we'll start using them before SDCC 4.6.0. and iy, de call (hl) call (iy) cbm #n clr hl cp hl, #d cp hl, de dwjnz label ex bc, hl flag cc, hl ; should this be flag cc, hl or flag hl, cc? ibox a jp cx,...
-
Updated list of instruction based on work in Rabbit code generation done since. For most of these instructions, we have now prepared support in r4k codegen, so we could start testing them easily. For the few in the list, where codegen isn't prepared yet, I included them since I consider it likely that we'll start using them before SDCC 4.6.0. and iy, de call (hl) call (iy) cbm #n clr hl cp hl, #d cp hl, de dwjnz label ex bc, hl flag cc, hl ; should this be flag cc, hl or flag hl, cc? ibox a jp cx,...
-
Updated list of instruction based on work in Rabbit code generation done since. For most of these instructions, we have now prepared support in r4k codegen, so we could start testing them easily. For the few in the list, where codegen isn't prepared yet, I included them since I consider it likely that we'll start using them before SDCC 4.6.0. and iy, de call (hl) call (iy) cbm #n clr hl cp hl, #d cp hl, de dwjnz label ex bc, hl flag cc, hl ; should this be flag cc, hl or flag hl, cc? ibox a jp cx,...
-
The current assembler has some r4k support, but only for R4K/11 mode. I think we need major changes in the assembler. This summarizes how IMO it should work: .r2k - Just the Rabbit 2000/3000 instructions .r3ka - Enables support for the additional Rabbit 3000A instructions .r4k_00 - Enables support for the additional Rabbit 4000 instructions available in mode 00 (aka default instruction set aka Rabbit 2000/3000 instruction set) .r4k_11 - Assign opcodes according to mode 11 (aka enhanced instruction...
-
Merges from trunk.
-
Merge pqc branch.
-
Try to fix mergeinfo on .version
-
Merges from trunk.
-
Prepare for future rotation codegen improvement for r4k/r5k/r6k port.
-
Fix -Wwrite-strings warning
-
Make plain int bit-fields signed.
-
Fix missing non-inline definition in a regression test.
-
Fix missing non-inline definition in a regression test.
-
Fix stdbit-h test-host issue for Ascon and Serpent tests.
-
Fix the names of mixed-sign support functions for multiplicative operators.
-
ChangeLog entry for pqc branch merge
-
Merge pqc branch.
-
Regression test for Falcon PQC NTT.
-
Regression test for Dilithium PQC NTT.
-
Regression test for Kyber PQC NTT.
-
Add regression test for sixth implementation of the Ascon-AEAD128 cipher and the Ascon-Hash256 hash function.
-
Merges from trunk.
-
Expose prototypes of fixed-point math support functions for pic16 only, since no other port supports fixed-point math.
-
More fine-grained Ascon test enabling for ports.
-
Fix compilation failure part of bug #3920, and a related issue affecting compound literals in inlined functions.
-
Merges from trunk.
-
Merge pqc branch.
-
Merges from upstream branches.
-
Add r4k Coremark graphs.
-
Add regression test for fifth implementation of the Ascon-AEAD128 cipher and the Ascon-Hash256 hash function.
-
Merges from trunk.
-
Fix bug #3712.
-
Regression tests for four implementations of the Ascon-AEAD128 cipher and the Ascon-Hash256 hash function each.
-
Merges from trunk.
-
Merge _Optional branch.
-
Merges from trunk.
-
Fix a regression in warnings for uninitalized variables.
1 >
