* [libre-riscv-dev] [Bug 342] formal proof of soc.fu.compunits.FunctionUnitBaseSingle needed
From: bugzilla-daemon @ 2020-05-24 18:16 UTC (permalink / raw)
To: libre-riscv-dev
In-Reply-To: <bug-342-13@https.bugs.libre-soc.org/>
https://bugs.libre-soc.org/show_bug.cgi?id=342
Michael Nolan <mtnolan2640@gmail.com> changed:
What |Removed |Added
----------------------------------------------------------------------------
CC| |mtnolan2640@gmail.com
Assignee|lkcl@lkcl.net |mtnolan2640@gmail.com
--
You are receiving this mail because:
You are on the CC list for the bug.
_______________________________________________
libre-riscv-dev mailing list
libre-riscv-dev@lists.libre-riscv.org
http://lists.libre-riscv.org/mailman/listinfo/libre-riscv-dev
^ permalink raw reply
* [libre-riscv-dev] [Bug 316] bperm TODO
From: bugzilla-daemon @ 2020-05-24 18:13 UTC (permalink / raw)
To: libre-riscv-dev
In-Reply-To: <bug-316-13@https.bugs.libre-soc.org/>
https://bugs.libre-soc.org/show_bug.cgi?id=316
--- Comment #97 from Michael Nolan <mtnolan2640@gmail.com> ---
(In reply to Luke Kenneth Casson Leighton from comment #96)
> (In reply to Michael Nolan from comment #95)
> > (In reply to Cole Poirier from comment #94)
> > > Michael, the unit tests for fu/logical/test/test_pipe_caller.py has been
> > > fixed by luke, and now produces a failed test for my module bpermd. Thanks
> > > for your help!
> >
> > Bpermd is now working
>
> *smacks-forehead* - the commit you made, to use self.i.b, made no sense,
> until
> i spotted in the popcount for-loop that b had been overwritten. that's why
> it ended up with the value "7".
>
> and the horrible thing is, the simulation would only have failed after
> running popcnt before bpermd.
Oh is *THAT* why? That confused the heck out of me.
--
You are receiving this mail because:
You are on the CC list for the bug.
_______________________________________________
libre-riscv-dev mailing list
libre-riscv-dev@lists.libre-riscv.org
http://lists.libre-riscv.org/mailman/listinfo/libre-riscv-dev
^ permalink raw reply
* [libre-riscv-dev] [Bug 316] bperm TODO
From: bugzilla-daemon @ 2020-05-24 17:57 UTC (permalink / raw)
To: libre-riscv-dev
In-Reply-To: <bug-316-13@https.bugs.libre-soc.org/>
https://bugs.libre-soc.org/show_bug.cgi?id=316
--- Comment #96 from Luke Kenneth Casson Leighton <lkcl@lkcl.net> ---
(In reply to Michael Nolan from comment #95)
> (In reply to Cole Poirier from comment #94)
> > Michael, the unit tests for fu/logical/test/test_pipe_caller.py has been
> > fixed by luke, and now produces a failed test for my module bpermd. Thanks
> > for your help!
>
> Bpermd is now working
*smacks-forehead* - the commit you made, to use self.i.b, made no sense, until
i spotted in the popcount for-loop that b had been overwritten. that's why
it ended up with the value "7".
and the horrible thing is, the simulation would only have failed after
running popcnt before bpermd.
--- a/src/soc/fu/logical/main_stage.py
+++ b/src/soc/fu/logical/main_stage.py
@@ -45,7 +45,6 @@ class LogicalMainStage(PipeModBase):
comb += o.ok.eq(1) # overridden if no op activates
-
m.submodules.bpermd = bpermd = Bpermd(64)
##########################
@@ -76,13 +75,13 @@ class LogicalMainStage(PipeModBase):
pc = [a]
# QTY32 2-bit (to take 2x 1-bit sums) etc.
work = [(32, 2), (16, 3), (8, 4), (4, 5), (2, 6), (1, 7)]
- for l, b in work:
- pc.append(array_of(l, b))
+ for l, bw in work:
+ pc.append(array_of(l, bw))
pc8 = pc[3] # array of 8 8-bit counts (popcntb)
pc32 = pc[5] # array of 2 32-bit counts (popcntw)
popcnt = pc[-1] # array of 1 64-bit count (popcntd)
# cascade-tree of adds
- for idx, (l, b) in enumerate(work):
+ for idx, (l, bw) in enumerate(work):
for i in range(l):
stt, end = i*2, i*2+1
src, dst = pc[idx], pc[idx+1]
@@ -136,7 +135,7 @@ class LogicalMainStage(PipeModBase):
###### bpermd #######
with m.Case(InternalOp.OP_BPERM):
comb += bpermd.rs.eq(a)
- comb += bpermd.rb.eq(self.i.b)
+ comb += bpermd.rb.eq(b)
comb += o.data.eq(bpermd.ra)
with m.Default():
--
You are receiving this mail because:
You are on the CC list for the bug.
_______________________________________________
libre-riscv-dev mailing list
libre-riscv-dev@lists.libre-riscv.org
http://lists.libre-riscv.org/mailman/listinfo/libre-riscv-dev
^ permalink raw reply
* [libre-riscv-dev] [Bug 316] bperm TODO
From: bugzilla-daemon @ 2020-05-24 16:45 UTC (permalink / raw)
To: libre-riscv-dev
In-Reply-To: <bug-316-13@https.bugs.libre-soc.org/>
https://bugs.libre-soc.org/show_bug.cgi?id=316
--- Comment #95 from Michael Nolan <mtnolan2640@gmail.com> ---
(In reply to Cole Poirier from comment #94)
> Michael, the unit tests for fu/logical/test/test_pipe_caller.py has been
> fixed by luke, and now produces a failed test for my module bpermd. Thanks
> for your help!
Bpermd is now working
--
You are receiving this mail because:
You are on the CC list for the bug.
_______________________________________________
libre-riscv-dev mailing list
libre-riscv-dev@lists.libre-riscv.org
http://lists.libre-riscv.org/mailman/listinfo/libre-riscv-dev
^ permalink raw reply
* [libre-riscv-dev] [Bug 346] New: simplified test link between compunits and regfile
From: bugzilla-daemon @ 2020-05-24 15:33 UTC (permalink / raw)
To: libre-riscv-dev
https://bugs.libre-soc.org/show_bug.cgi?id=346
Bug ID: 346
Summary: simplified test link between compunits and regfile
Product: Libre-SOC's first SoC
Version: unspecified
Hardware: PC
OS: Mac OS
Status: CONFIRMED
Severity: enhancement
Priority: ---
Component: Source Code
Assignee: lkcl@lkcl.net
Reporter: lkcl@lkcl.net
CC: libre-riscv-dev@lists.libre-riscv.org
NLnet milestone: ---
see
http://lists.libre-riscv.org/pipermail/libre-riscv-dev/2020-May/007257.html
--
You are receiving this mail because:
You are on the CC list for the bug.
_______________________________________________
libre-riscv-dev mailing list
libre-riscv-dev@lists.libre-riscv.org
http://lists.libre-riscv.org/mailman/listinfo/libre-riscv-dev
^ permalink raw reply
* [libre-riscv-dev] idea for testing pipelines
From: Luke Kenneth Casson Leighton @ 2020-05-24 15:30 UTC (permalink / raw)
To: Libre-RISCV General Development
i have an idea for testing the pipelines and also the regfiles, which
could also be a (drastically simpler, brain-dead, very slow)
processor:
* to create a system which, by monitoring CompUnit "busy" signals,
only permits one pipeline at a time to be active.
* to set up some very VERY simple connectivity between CompUnits and
Regfiles, involving either a single broadcast bus or a full crossbar
that would include incrementing the PC, so as to be able to do
instruction sequences.
this would allow us to confirm the functionality of the pipelines,
without needing the Dependency Matrices, which is a big piece of HDL
and will run really slowly in simulation.
l.
---
crowd-funded eco-conscious hardware: https://www.crowdsupply.com/eoma68
_______________________________________________
libre-riscv-dev mailing list
libre-riscv-dev@lists.libre-riscv.org
http://lists.libre-riscv.org/mailman/listinfo/libre-riscv-dev
^ permalink raw reply
* [libre-riscv-dev] [Bug 336] ALU CompUnit needs to recognise that RA (src1) can be zero
From: bugzilla-daemon @ 2020-05-24 15:07 UTC (permalink / raw)
To: libre-riscv-dev
In-Reply-To: <bug-336-13@https.bugs.libre-soc.org/>
https://bugs.libre-soc.org/show_bug.cgi?id=336
--- Comment #17 from Luke Kenneth Casson Leighton <lkcl@lkcl.net> ---
(In reply to Luke Kenneth Casson Leighton from comment #16)
> sure. i'd suggest doing this:
> latchregister(m, self.get_out(i), data_r, req_l.q[i], name)
> -->
> latchregister(m, self.get_out(i), data_r, req_l.q[i], name+"_l")
> sure - feel free to call it "oper_l" instead
btw, commit this immediately - do not wait to "accumulate" a series of commits,
and *definitely* do not put multiple separate and distinct "purposes" into
the one commit.
git is intelligent enough to merge commits that are not on the exact same line.
therefore, to speed up collaboration, do take advantage of that: do not wait,
just commit it - and commit "single-purpose". if you find you want to use the
word "and" in the commit message, *stop* - it means that the commit is not
single-purpose. more on that in HDL_workflow.
--
You are receiving this mail because:
You are on the CC list for the bug.
_______________________________________________
libre-riscv-dev mailing list
libre-riscv-dev@lists.libre-riscv.org
http://lists.libre-riscv.org/mailman/listinfo/libre-riscv-dev
^ permalink raw reply
* [libre-riscv-dev] [Bug 345] New: define POWER9 regfiles
From: bugzilla-daemon @ 2020-05-24 14:55 UTC (permalink / raw)
To: libre-riscv-dev
https://bugs.libre-soc.org/show_bug.cgi?id=345
Bug ID: 345
Summary: define POWER9 regfiles
Product: Libre-SOC's first SoC
Version: unspecified
Hardware: PC
OS: Mac OS
Status: CONFIRMED
Severity: enhancement
Priority: ---
Component: Source Code
Assignee: lkcl@lkcl.net
Reporter: lkcl@lkcl.net
CC: libre-riscv-dev@lists.libre-riscv.org
NLnet milestone: ---
a series of regfiles are needed, with different characteristics and a common
API
--
You are receiving this mail because:
You are on the CC list for the bug.
_______________________________________________
libre-riscv-dev mailing list
libre-riscv-dev@lists.libre-riscv.org
http://lists.libre-riscv.org/mailman/listinfo/libre-riscv-dev
^ permalink raw reply
* [libre-riscv-dev] [Bug 336] ALU CompUnit needs to recognise that RA (src1) can be zero
From: bugzilla-daemon @ 2020-05-24 14:44 UTC (permalink / raw)
To: libre-riscv-dev
In-Reply-To: <bug-336-13@https.bugs.libre-soc.org/>
https://bugs.libre-soc.org/show_bug.cgi?id=336
--- Comment #16 from Luke Kenneth Casson Leighton <lkcl@lkcl.net> ---
(In reply to Cesar Strauss from comment #15)
> Found this:
>
> https://git.libre-soc.org/?p=soc.git;a=blob;f=src/soc/experiment/
> compalu_multi.py;h=75df2153517ecd733cd5270776e88d76b8df5d06;hb=HEAD#l213
>
> name = "data_r%d" % i
> data_r = Signal(self.cu._get_srcwid(i), name=name, reset_less=True)
> latchregister(m, self.get_out(i), data_r, req_l.q[i], name)
>
> We are giving the same name for two things, the output signal of the
> latchregister, and its internal DFF (given by the function parameter).
> nMigen will append "$1" to one of them, but which?
wonderful, isn't it?
> Can we go and choose a name, given to its internal DFF, different from its
> output, in this and similar cases?
sure. i'd suggest doing this:
latchregister(m, self.get_out(i), data_r, req_l.q[i], name)
-->
latchregister(m, self.get_out(i), data_r, req_l.q[i], name+"_l")
>
> Also:
>
> https://git.libre-soc.org/?p=soc.git;a=blob;f=src/soc/experiment/
> compalu_multi.py;h=75df2153517ecd733cd5270776e88d76b8df5d06;hb=HEAD#l208
>
> latchregister(m, self.oper_i, oper_r, self.issue_i, "oper_r")
sure - feel free to call it "oper_l" instead
--
You are receiving this mail because:
You are on the CC list for the bug.
_______________________________________________
libre-riscv-dev mailing list
libre-riscv-dev@lists.libre-riscv.org
http://lists.libre-riscv.org/mailman/listinfo/libre-riscv-dev
^ permalink raw reply
* [libre-riscv-dev] [Bug 336] ALU CompUnit needs to recognise that RA (src1) can be zero
From: bugzilla-daemon @ 2020-05-24 14:37 UTC (permalink / raw)
To: libre-riscv-dev
In-Reply-To: <bug-336-13@https.bugs.libre-soc.org/>
https://bugs.libre-soc.org/show_bug.cgi?id=336
--- Comment #15 from Cesar Strauss <cestrauss@gmail.com> ---
Found this:
https://git.libre-soc.org/?p=soc.git;a=blob;f=src/soc/experiment/compalu_multi.py;h=75df2153517ecd733cd5270776e88d76b8df5d06;hb=HEAD#l213
name = "data_r%d" % i
data_r = Signal(self.cu._get_srcwid(i), name=name, reset_less=True)
latchregister(m, self.get_out(i), data_r, req_l.q[i], name)
We are giving the same name for two things, the output signal of the
latchregister, and its internal DFF (given by the function parameter). nMigen
will append "$1" to one of them, but which?
Can we go and choose a name, given to its internal DFF, different from its
output, in this and similar cases?
Also:
https://git.libre-soc.org/?p=soc.git;a=blob;f=src/soc/experiment/compalu_multi.py;h=75df2153517ecd733cd5270776e88d76b8df5d06;hb=HEAD#l208
latchregister(m, self.oper_i, oper_r, self.issue_i, "oper_r")
In this case, as it happens, oper_r sub-signals do not receive names that begin
with "oper_r". So, what receives the "oper_r" name is actually the internal
DFF, which confused me at no end on the GTKWave timing diagram. This is how I
found this issue.
--
You are receiving this mail because:
You are on the CC list for the bug.
_______________________________________________
libre-riscv-dev mailing list
libre-riscv-dev@lists.libre-riscv.org
http://lists.libre-riscv.org/mailman/listinfo/libre-riscv-dev
^ permalink raw reply
* [libre-riscv-dev] [Bug 344] New: missing mtmsr and mfsprd
From: bugzilla-daemon @ 2020-05-24 14:27 UTC (permalink / raw)
To: libre-riscv-dev
https://bugs.libre-soc.org/show_bug.cgi?id=344
Bug ID: 344
Summary: missing mtmsr and mfsprd
Product: Libre-SOC's first SoC
Version: unspecified
Hardware: PC
OS: Mac OS
Status: CONFIRMED
Severity: enhancement
Priority: ---
Component: Source Code
Assignee: lkcl@lkcl.net
Reporter: lkcl@lkcl.net
CC: libre-riscv-dev@lists.libre-riscv.org
NLnet milestone: ---
missing two operations from sprset.mdwn and also minor_op31.csv
Move To Machine State Register
X-form
mtmsr RS,L
if L = 0 then
MSR48 ← (RS)48 | (RS)49
MSR58 <- ((RS)58 | (RS)49)& ¬(MSR41 & MSR3 & (¬(RS)49))
MSR59 <- ((RS)59 | (RS)49)& ¬(MSR41 & MSR3 & (¬(RS)49))
MSR32:40 42:47 49:50 52:57 60:62 <- (RS)32:40 42:47 49:50 52:57 60:62
else
MSR48 62 ← (RS)48 62
Move To Machine State Register Doubleword
X-Form
mtmsrd RS,L
if L = 0 then
MSR48 ← (RS)48 | (RS)49
MSR58 <- ((RS)58 | (RS)49)& ¬(MSR41 & MSR3 & (¬(RS)49))
MSR59 <- ((RS)59 | (RS)49) & ¬(MSR41 & MSR3 & (¬(RS)49))
MSR0:2 4:40 42:47 49:50 52:57 60:62 ← (RS)0:2 4 6:40 42:47 49:50 52:57 60:62
else
MSR48 62 ← (RS)48 62
--
You are receiving this mail because:
You are on the CC list for the bug.
_______________________________________________
libre-riscv-dev mailing list
libre-riscv-dev@lists.libre-riscv.org
http://lists.libre-riscv.org/mailman/listinfo/libre-riscv-dev
^ permalink raw reply
* [libre-riscv-dev] daily kan-ban update 24may2020
From: Luke Kenneth Casson Leighton @ 2020-05-24 14:17 UTC (permalink / raw)
To: Libre-RISCV General Development
yesterday:
* morphed MultiCompUnit to use regspec API
* wrote a module that understands the regspec API and documented it
* created a stub DataMerger class for Tobias to implement
* helped Cesar prioritise tasks
* created soc.fu.compunits with help from Cole
* created some more issues to keep track of forward-planned upcoming work
* reviewed pipeline code
* noticed that we can add OP_MT/FMSR to OP_TRAP
this morning:
* updated the pipelines, tests and formal proofs to use Data for INT regs.
* put a bugracker link to everyone's assignments on their wik homepage
today:
* possibly create a series of register files (or at least the bugreport for it)
* plan some unit test infrastructure for soc.fu.compunits
* split out unit tests from all test_pipe_caller.py and look at using
it for testing soc.fu.compunits
_______________________________________________
libre-riscv-dev mailing list
libre-riscv-dev@lists.libre-riscv.org
http://lists.libre-riscv.org/mailman/listinfo/libre-riscv-dev
^ permalink raw reply
* [libre-riscv-dev] [Bug 336] ALU CompUnit needs to recognise that RA (src1) can be zero
From: bugzilla-daemon @ 2020-05-24 14:00 UTC (permalink / raw)
To: libre-riscv-dev
In-Reply-To: <bug-336-13@https.bugs.libre-soc.org/>
https://bugs.libre-soc.org/show_bug.cgi?id=336
--- Comment #14 from Luke Kenneth Casson Leighton <lkcl@lkcl.net> ---
(In reply to Cesar Strauss from comment #13)
> Still need a way to test the case of zero_a and/or imm_data being entirely
> absent, in which case the muxes are not generated.
ah if you create a test which uses... mmm.... which one is dead-simple...
CompCROpSubset, then that doesn't have imm_data or zero_a.
see soc.experiment.alu_hier.DummyALU
i've just added a DummyALU for you, it ignores operand b, ignores the
actual operand type (up to you if you want to put that back, do something
different), just copies a to o - nothing sophisticated.
so you can copy test_compunit_regspec1(), call it test_compunit_regspec2()
then cookie-cut scoreboard_sim() and op_sim() to remove trying to set imm_data,
invert_a, and zero_a (because those don't exist in CompCROpSubset), and
just assert that the result is equal to the 1st operand, a.
if you really wanted to you could actually remove b from the inspec in
test_compunit_regspec2() and also in DummyALU.
--
You are receiving this mail because:
You are on the CC list for the bug.
_______________________________________________
libre-riscv-dev mailing list
libre-riscv-dev@lists.libre-riscv.org
http://lists.libre-riscv.org/mailman/listinfo/libre-riscv-dev
^ permalink raw reply
* [libre-riscv-dev] [Bug 336] ALU CompUnit needs to recognise that RA (src1) can be zero
From: bugzilla-daemon @ 2020-05-24 13:46 UTC (permalink / raw)
To: libre-riscv-dev
In-Reply-To: <bug-336-13@https.bugs.libre-soc.org/>
https://bugs.libre-soc.org/show_bug.cgi?id=336
--- Comment #13 from Cesar Strauss <cestrauss@gmail.com> ---
(In reply to Luke Kenneth Casson Leighton from comment #11)
> brilliant, just saw this, Cesar. confirmed that it is functional.
>
> commit 260625df9309f8f35541207cab431dd8dba90c5a
> Author: Cesar Strauss <cestrauss@gmail.com>
> Date: Sat May 23 19:52:08 2020 -0300
>
> Add a few test cases with zero_a set, in combination with imm_ok
>
Still need a way to test the case of zero_a and/or imm_data being entirely
absent, in which case the muxes are not generated.
--
You are receiving this mail because:
You are on the CC list for the bug.
_______________________________________________
libre-riscv-dev mailing list
libre-riscv-dev@lists.libre-riscv.org
http://lists.libre-riscv.org/mailman/listinfo/libre-riscv-dev
^ permalink raw reply
* [libre-riscv-dev] [Bug 314] Create POWER9 Condition Register pipeline
From: bugzilla-daemon @ 2020-05-24 13:11 UTC (permalink / raw)
To: libre-riscv-dev
In-Reply-To: <bug-314-13@https.bugs.libre-soc.org/>
https://bugs.libre-soc.org/show_bug.cgi?id=314
--- Comment #21 from Luke Kenneth Casson Leighton <lkcl@lkcl.net> ---
unit test is not testing reg output (MFCRF, ISEL)
--
You are receiving this mail because:
You are on the CC list for the bug.
_______________________________________________
libre-riscv-dev mailing list
libre-riscv-dev@lists.libre-riscv.org
http://lists.libre-riscv.org/mailman/listinfo/libre-riscv-dev
^ permalink raw reply
* [libre-riscv-dev] [Bug 337] Convention for register outputs in *OutputData structures is to use "Data"
From: bugzilla-daemon @ 2020-05-24 12:11 UTC (permalink / raw)
To: libre-riscv-dev
In-Reply-To: <bug-337-13@https.bugs.libre-soc.org/>
https://bugs.libre-soc.org/show_bug.cgi?id=337
Luke Kenneth Casson Leighton <lkcl@lkcl.net> changed:
What |Removed |Added
----------------------------------------------------------------------------
Assignee|lkcl@lkcl.net |colepoirier@gmail.com
--
You are receiving this mail because:
You are on the CC list for the bug.
_______________________________________________
libre-riscv-dev mailing list
libre-riscv-dev@lists.libre-riscv.org
http://lists.libre-riscv.org/mailman/listinfo/libre-riscv-dev
^ permalink raw reply
* [libre-riscv-dev] [Bug 325] create POWER9 TRAP pipeline
From: bugzilla-daemon @ 2020-05-24 3:43 UTC (permalink / raw)
To: libre-riscv-dev
In-Reply-To: <bug-325-13@https.bugs.libre-soc.org/>
https://bugs.libre-soc.org/show_bug.cgi?id=325
--- Comment #10 from Luke Kenneth Casson Leighton <lkcl@lkcl.net> ---
could also add SPRSET as well (1 more SPR reg to input)
--
You are receiving this mail because:
You are on the CC list for the bug.
_______________________________________________
libre-riscv-dev mailing list
libre-riscv-dev@lists.libre-riscv.org
http://lists.libre-riscv.org/mailman/listinfo/libre-riscv-dev
^ permalink raw reply
* [libre-riscv-dev] [Bug 325] create POWER9 TRAP pipeline
From: bugzilla-daemon @ 2020-05-24 3:41 UTC (permalink / raw)
To: libre-riscv-dev
In-Reply-To: <bug-325-13@https.bugs.libre-soc.org/>
https://bugs.libre-soc.org/show_bug.cgi?id=325
--- Comment #9 from Luke Kenneth Casson Leighton <lkcl@lkcl.net> ---
suggest adding mfmsr and mtmsr to TRAP pipeline. does mean adding RT and MSR
as additional output regs.
--
You are receiving this mail because:
You are on the CC list for the bug.
_______________________________________________
libre-riscv-dev mailing list
libre-riscv-dev@lists.libre-riscv.org
http://lists.libre-riscv.org/mailman/listinfo/libre-riscv-dev
^ permalink raw reply
* [libre-riscv-dev] [Bug 336] ALU CompUnit needs to recognise that RA (src1) can be zero
From: bugzilla-daemon @ 2020-05-24 3:33 UTC (permalink / raw)
To: libre-riscv-dev
In-Reply-To: <bug-336-13@https.bugs.libre-soc.org/>
https://bugs.libre-soc.org/show_bug.cgi?id=336
Luke Kenneth Casson Leighton <lkcl@lkcl.net> changed:
What |Removed |Added
----------------------------------------------------------------------------
See Also| |https://bugs.libre-soc.org/
| |show_bug.cgi?id=343
--
You are receiving this mail because:
You are on the CC list for the bug.
_______________________________________________
libre-riscv-dev mailing list
libre-riscv-dev@lists.libre-riscv.org
http://lists.libre-riscv.org/mailman/listinfo/libre-riscv-dev
^ permalink raw reply
* [libre-riscv-dev] [Bug 343] compalu_multi write requests need to hook into Data.ok
From: bugzilla-daemon @ 2020-05-24 3:33 UTC (permalink / raw)
To: libre-riscv-dev
In-Reply-To: <bug-343-13@https.bugs.libre-soc.org/>
https://bugs.libre-soc.org/show_bug.cgi?id=343
Luke Kenneth Casson Leighton <lkcl@lkcl.net> changed:
What |Removed |Added
----------------------------------------------------------------------------
NLnet milestone|--- |NLnet.2019.02
See Also| |https://bugs.libre-soc.org/
| |show_bug.cgi?id=336
--
You are receiving this mail because:
You are on the CC list for the bug.
_______________________________________________
libre-riscv-dev mailing list
libre-riscv-dev@lists.libre-riscv.org
http://lists.libre-riscv.org/mailman/listinfo/libre-riscv-dev
^ permalink raw reply
* [libre-riscv-dev] [Bug 343] New: compalu_multi write requests need to hook into Data.ok
From: bugzilla-daemon @ 2020-05-24 3:33 UTC (permalink / raw)
To: libre-riscv-dev
https://bugs.libre-soc.org/show_bug.cgi?id=343
Bug ID: 343
Summary: compalu_multi write requests need to hook into Data.ok
Product: Libre-SOC's first SoC
Version: unspecified
Hardware: Other
OS: Linux
Status: CONFIRMED
Severity: enhancement
Priority: ---
Component: Source Code
Assignee: lkcl@lkcl.net
Reporter: lkcl@lkcl.net
CC: libre-riscv-dev@lists.libre-riscv.org
NLnet milestone: ---
the soc.fu pipeline data structures have been designed to notify MultiCompUnit
that they have data to be stored.
this by setting Data.ok in their Output n.data_o members.
--
You are receiving this mail because:
You are on the CC list for the bug.
_______________________________________________
libre-riscv-dev mailing list
libre-riscv-dev@lists.libre-riscv.org
http://lists.libre-riscv.org/mailman/listinfo/libre-riscv-dev
^ permalink raw reply
* [libre-riscv-dev] [Bug 336] ALU CompUnit needs to recognise that RA (src1) can be zero
From: bugzilla-daemon @ 2020-05-24 1:46 UTC (permalink / raw)
To: libre-riscv-dev
In-Reply-To: <bug-336-13@https.bugs.libre-soc.org/>
https://bugs.libre-soc.org/show_bug.cgi?id=336
--- Comment #12 from Luke Kenneth Casson Leighton <lkcl@lkcl.net> ---
https://git.libre-soc.org/?p=nmutil.git;a=blob;f=src/nmutil/test/test_buf_pipe.py;hb=HEAD
just occurred to me, see class Test3, this nay be simpler to start from.
remove the for loops for now, add a parameter read_idx to the send function,
and write_idx to rcv function.
read_idx will refer to self.compunit.rd.rel[read_idx] you see how that would
go?
--
You are receiving this mail because:
You are on the CC list for the bug.
_______________________________________________
libre-riscv-dev mailing list
libre-riscv-dev@lists.libre-riscv.org
http://lists.libre-riscv.org/mailman/listinfo/libre-riscv-dev
^ permalink raw reply
* [libre-riscv-dev] Fwd: [OpenPOWER-HDL-Cores] POWER9 ISA opcode usage statistics
From: Luke Kenneth Casson Leighton @ 2020-05-24 1:01 UTC (permalink / raw)
To: Libre-RISCV General Development
In-Reply-To: <CAO3ALPwjLPh8y_UAucNxZ0Zc9TjgQZmPgTq8M4fuQLhiUioThg@mail.gmail.com>
can someone please capture this in a bugreport?
---
crowd-funded eco-conscious hardware: https://www.crowdsupply.com/eoma68
---------- Forwarded message ---------
From: Forest Crossman <cyrozap@gmail.com>
Date: Sun, May 24, 2020 at 12:31 AM
Subject: Re: [OpenPOWER-HDL-Cores] POWER9 ISA opcode usage statistics
To: Luke Kenneth Casson Leighton <lkcl@lkcl.net>
Cc: <openpower-hdl-cores@mailinglist.openpowerfoundation.org>
> whilst i appreciate that these statistics will be use-case-specific
> for certain workloads, does anyone know of any studies that have
> already been done, or of any libre-licensed tools that help with the
> same?
I realize it's been over three weeks since you posted this question,
so maybe you've found an answer already, but I had an idea on how to
do it so I decided to try it. I ended up with the following bash
script: https://paste.debian.net/hidden/10201ca1
I ran that script on my Talos II running Debian 8 and got the
following results: https://paste.debian.net/hidden/13696b59
Note that Debian for ppc64el is built without `-mcpu=power9`, so these
statistics won't include any POWER9 instructions. If you want to get
the stats for that, you'll want to build Debian (or another popular
distro) with the `-mcpu=power9` flag to make sure the POWER9
instructions and optimizations are enabled. It also might be a good
idea to gather instruction statistics for the kernel as well, since
kernel code might have a different distribution than the userspace
code I sampled.
_______________________________________________
libre-riscv-dev mailing list
libre-riscv-dev@lists.libre-riscv.org
http://lists.libre-riscv.org/mailman/listinfo/libre-riscv-dev
^ permalink raw reply
* [libre-riscv-dev] [Bug 342] formal proof of soc.fu.compunits.FunctionUnitBaseSingle needed
From: bugzilla-daemon @ 2020-05-23 23:17 UTC (permalink / raw)
To: libre-riscv-dev
In-Reply-To: <bug-342-13@https.bugs.libre-soc.org/>
https://bugs.libre-soc.org/show_bug.cgi?id=342
--- Comment #1 from Luke Kenneth Casson Leighton <lkcl@lkcl.net> ---
the tricky bit: MultiCompUnit is time-dependent. that involves use
of "Past()". it is similar in some ways to the formal proof for
SyncFIFO because it is similar signalling rules on each of the
REQ/GO signals.
--
You are receiving this mail because:
You are on the CC list for the bug.
_______________________________________________
libre-riscv-dev mailing list
libre-riscv-dev@lists.libre-riscv.org
http://lists.libre-riscv.org/mailman/listinfo/libre-riscv-dev
^ permalink raw reply
* [libre-riscv-dev] [Bug 336] ALU CompUnit needs to recognise that RA (src1) can be zero
From: bugzilla-daemon @ 2020-05-23 23:10 UTC (permalink / raw)
To: libre-riscv-dev
In-Reply-To: <bug-336-13@https.bugs.libre-soc.org/>
https://bugs.libre-soc.org/show_bug.cgi?id=336
--- Comment #11 from Luke Kenneth Casson Leighton <lkcl@lkcl.net> ---
brilliant, just saw this, Cesar. confirmed that it is functional.
commit 260625df9309f8f35541207cab431dd8dba90c5a
Author: Cesar Strauss <cestrauss@gmail.com>
Date: Sat May 23 19:52:08 2020 -0300
Add a few test cases with zero_a set, in combination with imm_ok
so the next one is: we need something that's a bit more sophisticated,
designed more along the lines of test_inout_mux_pipe.py:
https://git.libre-soc.org/?p=nmutil.git;a=blob;f=src/nmutil/test/test_inout_mux_pipe.py;h=03fddde0f7ab12ec9023fc57b913442ba42c120a;hb=HEAD#l52
not quite that sophisticated, but close.
basically, we need:
* one function (similar to TestInput.send) which takes care of *one*
REQ_READ and GO_RD signalling bit. it should monitor:
self.rd.req[N], then set self.rd.go[N] for *one* cycle (one blank yield)
then set it self.rd.go[N] back to zero and confirm (assert) that
self.rd.req[N] has gone to zero
* another function (similar again to TestInput.rcv) which does the same
thing for REQ_WRITE and GO_WRITE.
* another function - a very simple one - which sets issue_i
for one cycle (yield in between), then waits for "self.busy" to drop to 0
we can make it more sophisticated as we go along.
it should be... only... about... 80 lines of code, at a guess.
--
You are receiving this mail because:
You are on the CC list for the bug.
_______________________________________________
libre-riscv-dev mailing list
libre-riscv-dev@lists.libre-riscv.org
http://lists.libre-riscv.org/mailman/listinfo/libre-riscv-dev
^ permalink raw reply
page: next prev latest
Libre RISC-V Mailing list Archives (accessible via public-inbox)
Archives are clonable: git clone --mirror https://inbox.libre-soc.org/libre-riscv-dev
Example config snippet for mirrors
AGPL code for this site: git clone https://public-inbox.org/public-inbox.git