OpenCores
URL https://opencores.org/ocsvn/xulalx25soc/xulalx25soc/trunk

Subversion Repositories xulalx25soc

[/] [xulalx25soc/] [trunk/] [rtl/] [cpu/] - Rev 115

Rev

Go to most recent revision

Filtering Options

Clear current filter

Rev Log message Author Age Path
113 Lots of updates, mostly focused on getting the XuLA board and CPU able to pass
a comprehensive test suite. At this point, everything passes and logic usage
is down even. Among the few changes includes the "break" bit in the uCC
register, used to indicate a switch to supervisor mode occurred as a result
of a user break, and the ability for the supervisor to clear the instruction
cache.
dgisselq 3042d 07h /xulalx25soc/trunk/rtl/cpu/
98 Updated copyright notices for the new year, to reflect that changes have been
made in 2016.
dgisselq 3051d 01h /xulalx25soc/trunk/rtl/cpu/
89 Bug fixes, following the adf_ce logic combining ALU/DIVIDE/FPU pipeline logic
into one register, this fixes that logic so that instructions without their
condition fulfilled are still "executed" and marked as done.
dgisselq 3074d 05h /xulalx25soc/trunk/rtl/cpu/
87 Placed the interrupt into the carry chain for less logic area. dgisselq 3074d 06h /xulalx25soc/trunk/rtl/cpu/
86 Fine tuning the `defines, so that you can build pipelined without pipelined
bus access and so forth.
dgisselq 3074d 06h /xulalx25soc/trunk/rtl/cpu/
73 Simplified logic. dgisselq 3080d 02h /xulalx25soc/trunk/rtl/cpu/
72 Sets XULA25 as the default. dgisselq 3080d 02h /xulalx25soc/trunk/rtl/cpu/
70 Cosmetic (minor) update. dgisselq 3080d 02h /xulalx25soc/trunk/rtl/cpu/
69 Massive logic simplification. This is also the first (verified) working
version.
dgisselq 3080d 02h /xulalx25soc/trunk/rtl/cpu/
68 Fixes the debug ack line, so it no longer acks when there isn't a dbg_stb.
Fixed the multiply option parameter, so it sets a 3-clock multiply properly.
Adjusted the watchdog timer so that it now produces a timer that doesn't
reload--since this is pointless for a watchdog. Finally, connects a reset
line to the DMA, to make certain that resetting the CPU will also reset any
ongoing DMA operation.
dgisselq 3080d 02h /xulalx25soc/trunk/rtl/cpu/
67 Simplifies logic, and guarantees that the minimum set value will always
produce an int. For example, if the count was X before, setting X-1 wouldn't
produce an interrupt since it had passed. Now it produces an interrupt, and
keeps the next interrupt valid.
dgisselq 3080d 02h /xulalx25soc/trunk/rtl/cpu/
66 Simplified logic (barely). dgisselq 3080d 02h /xulalx25soc/trunk/rtl/cpu/
65 Makes the auto-reload feature a parameterized (generic) feature, so the same
code will work for both auto-reloadable and non-autoreloadable (i.e. watchdog)
timers.
dgisselq 3080d 02h /xulalx25soc/trunk/rtl/cpu/
62 Removed the pipe logic from the non-pipelined version, and made the NOOP a
specific ALU instruction so that the PC is always properly updated.
dgisselq 3080d 02h /xulalx25soc/trunk/rtl/cpu/
61 Fixed the timing control wires: busy and valid will never both be true. Busy
will be true (now) until valid is asserted, and busy will never not be
asserted until valid is true.
dgisselq 3080d 02h /xulalx25soc/trunk/rtl/cpu/
60 LONG_MPY upgrade. This is part of swapping out the LDIHI instruction for a
MPY, and the MPYS and MPYU instructions for MPYSHI and MPYUHI respectively.
dgisselq 3080d 03h /xulalx25soc/trunk/rtl/cpu/
52 This brings the XuLA2-LX25 SoC up to speed with the rest of the ZipCPU, and
prepares it for the 32x32 bit multiply instruction set change.
dgisselq 3128d 03h /xulalx25soc/trunk/rtl/cpu/
51 Lots of bug fixes. The ugliest were in the prefetch cache, where instructions
from one cache line were being issued as valid in another. Other fixes include
pipeline fixes so that LOD (Rx),Rx; LOD(Rx),Rx works, and more. Finally, the
decode was adjusted so that brev no longer affects the flags.
dgisselq 3138d 01h /xulalx25soc/trunk/rtl/cpu/
50 Updates to fix some broken early branching code, both in idecode and pfcache. dgisselq 3147d 04h /xulalx25soc/trunk/rtl/cpu/
46 This is a bug fix release--fixing the bug that kept dumpsdram.cpp/wbsdram.v
from working when long pipelined reads were interrupted by the necessity of
a pair of refresh cycles.
dgisselq 3158d 06h /xulalx25soc/trunk/rtl/cpu/

1 2 Next >

Show All

powered by: WebSVN 2.1.0

© copyright 1999-2024 OpenCores.org, equivalent to Oliscience, all rights reserved. OpenCores®, registered trademark.