* [libre-riscv-dev] [Bug 320] split out software install and deps into separate script
From: bugzilla-daemon @ 2020-05-18 21:45 UTC (permalink / raw)
To: libre-riscv-dev
In-Reply-To: <bug-320-13@https.bugs.libre-soc.org/>
https://bugs.libre-soc.org/show_bug.cgi?id=320
--- Comment #11 from Luke Kenneth Casson Leighton <lkcl@lkcl.net> ---
superb. yes done tested therefore close. thank you cole.
--
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 320] split out software install and deps into separate script
From: bugzilla-daemon @ 2020-05-18 22:06 UTC (permalink / raw)
To: libre-riscv-dev
In-Reply-To: <bug-320-13@https.bugs.libre-soc.org/>
https://bugs.libre-soc.org/show_bug.cgi?id=320
Cole Poirier <colepoirier@gmail.com> changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|CONFIRMED |RESOLVED
Resolution|--- |FIXED
--
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 291] HDL Workflow and Coriolis2 chroot automated setup scripts
From: bugzilla-daemon @ 2020-05-18 22:15 UTC (permalink / raw)
To: libre-riscv-dev
In-Reply-To: <bug-291-13@https.bugs.libre-soc.org/>
https://bugs.libre-soc.org/show_bug.cgi?id=291
--- Comment #14 from Cole Poirier <colepoirier@gmail.com> ---
I have applied Jean-Paul's patch, committed and pushed it, and am now
continuing testing of the script.
--
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 @ 2020-05-18 22:29 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 #40 from Cole Poirier <colepoirier@gmail.com> ---
Hi Michael, I have finally run the formal proof 'proof_bperm.py' and gotten a
failure with the following in its output:
```
SBY 15:20:39 [proof_bperm_formal] engine_0: ## 0:00:00 BMC failed!
SBY 15:20:39 [proof_bperm_formal] engine_0: ## 0:00:00 Value for anyconst in
top (proof_bperm.py:56): 15481123719086080
SBY 15:20:39 [proof_bperm_formal] engine_0: ## 0:00:00 Value for anyconst in
top (proof_bperm.py:57): 36028797018963968
SBY 15:20:39 [proof_bperm_formal] engine_0: ## 0:00:00 Assert failed in top:
proof_bperm.py:93
```
I feel like I have misunderstood your directions for proving property #2, am I
extracting the indices incorrectly? Or am I making a different fundamental
mistake?
84 # Now we need to prove property #2. I'm going to leave this to
85 # you Cole. I'd start by writing a for loop and extracting the
86 # 8 indices into signals. Then I'd write an if statement
87 # checking if the index is >= 64 (it's hardware, so use an
88 # m.If()). Finally, I'd add an assert that checks whether
89 # ra[i] is equal to 0
90 for i in range(8):
91 index = rb[i*8:i*8+8]
92 with m.If(index >= 64):
93 comb += Assert(ra[i] == 0)
94
95
96 return m
--
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 @ 2020-05-18 22:36 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 #41 from Michael Nolan <mtnolan2640@gmail.com> ---
(In reply to Cole Poirier from comment #40)
>
> I feel like I have misunderstood your directions for proving property #2, am
> I extracting the indices incorrectly? Or am I making a different fundamental
> mistake?
>
> 84 # Now we need to prove property #2. I'm going to leave this to
> 85 # you Cole. I'd start by writing a for loop and extracting the
> 86 # 8 indices into signals. Then I'd write an if statement
> 87 # checking if the index is >= 64 (it's hardware, so use an
> 88 # m.If()). Finally, I'd add an assert that checks whether
> 89 # ra[i] is equal to 0
> 90 for i in range(8):
> 91 index = rb[i*8:i*8+8]
> 92 with m.If(index >= 64):
> 93 comb += Assert(ra[i] == 0)
> 94
> 95
> 96 return m
No, that is about what I would have written. What I would suggest to you is to
open the trace yosys generated (should be in
proof_bperm_formal/engine_0/trace.vcd) and see why it's not working.
Looking at it on my machine... double check that you're generating the `index`
correctly in the proof.
--
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 @ 2020-05-18 23:12 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 #42 from Cole Poirier <colepoirier@gmail.com> ---
(In reply to Michael Nolan from comment #41)
> No, that is about what I would have written. What I would suggest to you is
> to open the trace yosys generated (should be in
> proof_bperm_formal/engine_0/trace.vcd) and see why it's not working.
>
>
> Looking at it on my machine... double check that you're generating the
> `index` correctly in the proof.
When I look at trace.vcd by running 'gtkwave
proof_bperm_formal/engine_0/trace.vcd' and clicking to expose the bperm module
under top, then selecting and adding all the wires, all I see is, what I
presume out of ignorance is the initial state of the wires, and I can advance
to the next edge where the wires are in their 'final' state, and that's it...
I'm assuming since this is a combinatorial single cycle proof that's what I'm
supposed to see? How should I be reading the gtkwave diagram in order to
diagnose that index is not being set correctly in the proof?
Above, you said that doing 'index = rb[i*8:i*8+8]' within a for loop where i
counts from 0 to 7 was what you would have done... is this shown by the vcd
output to now be incorrect? How would I go about discerning that I'm generating
the `index` correctly in the proof? I'm sorry that these are stupid questions
that likely have obvious answers, but I'm struggling to determine them on my
own reading through the gtkwave manual.
--
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 @ 2020-05-18 23:17 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 #43 from Michael Nolan <mtnolan2640@gmail.com> ---
(In reply to Cole Poirier from comment #42)
> (In reply to Michael Nolan from comment #41)
> > No, that is about what I would have written. What I would suggest to you is
> > to open the trace yosys generated (should be in
> > proof_bperm_formal/engine_0/trace.vcd) and see why it's not working.
> >
> >
> > Looking at it on my machine... double check that you're generating the
> > `index` correctly in the proof.
>
> When I look at trace.vcd by running 'gtkwave
> proof_bperm_formal/engine_0/trace.vcd' and clicking to expose the bperm
> module under top, then selecting and adding all the wires, all I see is,
> what I presume out of ignorance is the initial state of the wires, and I can
> advance to the next edge where the wires are in their 'final' state, and
> that's it... I'm assuming since this is a combinatorial single cycle proof
> that's what I'm supposed to see? How should I be reading the gtkwave diagram
> in order to diagnose that index is not being set correctly in the proof?
For a combinatorial circuit like this one, the inputs (rs and rb) are set to
their initial values, and the outputs (ra) are set to whatever the circuit
calculates from the inputs.
>
> Above, you said that doing 'index = rb[i*8:i*8+8]' within a for loop where i
> counts from 0 to 7 was what you would have done... is this shown by the vcd
> output to now be incorrect? How would I go about discerning that I'm
> generating the `index` correctly in the proof? I'm sorry that these are
> stupid questions that likely have obvious answers, but I'm struggling to
> determine them on my own reading through the gtkwave manual.
In the proof, you're generating the index from rb, but in the actual module and
the Power spec, it's generated from rs. That's why the proof fails. If you look
at the index variables (for me it's index 6), you'll notice that one is less
than 64 - that's how I figured it out.
--
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 @ 2020-05-18 23:34 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 #44 from Cole Poirier <colepoirier@gmail.com> ---
(In reply to Michael Nolan from comment #43)
> (In reply to Cole Poirier from comment #42)
> For a combinatorial circuit like this one, the inputs (rs and rb) are set to
> their initial values, and the outputs (ra) are set to whatever the circuit
> calculates from the inputs.
Thank you for this clarification, I think I am missing some of the basic
background knowledge of registers/assembly, should I just do some background
reading on wikipedia, or is there a better resource you or luke can point me
to, or should I simply study OPF ISA Spec v3.1?
> In the proof, you're generating the index from rb, but in the actual module
> and the Power spec, it's generated from rs. That's why the proof fails. If
> you look at the index variables (for me it's index 6), you'll notice that
> one is less than 64 - that's how I figured it out.
That makes sense. I didn't catch this because all other idx_n's were 00, and
idx_6 was 37. 37 is less than 64, but is 00 not also less than 64 correct? I'm
confused by this so your clarification would be very helpful.
I'd like to clarify one other thing so I don't repeat this same mistake.
In the skeleton you provided the below pseudo code and assertion properties.
# The pseudocode in the Power ISA manual (v3.1) is as follows:
61 # do i = 0 to 7
62 # index <- RS[8*i:8*i+8]
63 # if index < 64:
64 # perm[i] <- RB[index]
65 # else:
66 # perm[i] <- 0
67 # RA <- 56'b0 || perm[0:8] # big endian though
68
69 # Looking at this, I can identify 3 properties that the bperm
70 # module should keep:
71 # 1. RA[8:64] should always equal 0
72 # 2. If RB[i*8:i*8+8] >= 64 then RA[i] should equal 0
73 # 3. If RB[i*8:i*8+8] < 64 then RA[i] should RS[index]
For 2 and 3 should RB[i*8:i*8+8] actually be RS[i*8:i*8+8]? Or am I failing to
properly read/interpret the 'formalism'? In other words, I'm very confused and
looking for a way to better understand what I am doing, and what I am supposed
to be doing.
--
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 @ 2020-05-18 23:54 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 #45 from Luke Kenneth Casson Leighton <lkcl@lkcl.net> ---
err *cough* really-bad-hack mumble *cough*
index 453e419..0f879e3 100644
--- a/src/soc/fu/logical/formal/proof_bperm.py
+++ b/src/soc/fu/logical/formal/proof_bperm.py
@@ -88,9 +88,15 @@ class Driver(Elaboratable):
# m.If()). Finally, I'd add an assert that checks whether
# ra[i] is equal to 0
for i in range(8):
- index = rb[i*8:i*8+8]
+ index = rs[i*8:i*8+8]
with m.If(index >= 64):
comb += Assert(ra[i] == 0)
+ with m.Else():
+ # to avoid having to create an Array of rb,
+ # cycle through from 0-63 on the index *whistle nonchalantly*
+ for j in range(64):
+ with m.If(index == j):
+ comb += Assert(ra[i] == rb[j])
return m
--
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 @ 2020-05-19 0:05 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 #46 from Luke Kenneth Casson Leighton <lkcl@lkcl.net> ---
(In reply to Cole Poirier from comment #44)
> 71 # 1. RA[8:64] should always equal 0
> 72 # 2. If RB[i*8:i*8+8] >= 64 then RA[i] should equal 0
> 73 # 3. If RB[i*8:i*8+8] < 64 then RA[i] should RS[index]
>
> For 2 and 3 should RB[i*8:i*8+8] actually be RS[i*8:i*8+8]?
yes.
> Or am I failing
> to properly read/interpret the 'formalism'? In other words, I'm very
> confused and looking for a way to better understand what I am doing, and
> what I am supposed to be doing.
you are expressing the expectations about the inputs and outputs of
the module.
these you express as assertions that must be true.
(sby then checks those for you under *all* possible values of input and
output).
therefore you not only need to get the code right, you actually need to
properly understand what it does... and express that in the assertions.
example.
from this line of code, is it *ever* possible that bits 8 and above
will be *anything* other than zero?
m.d.comb += self.ra[0:8].eq(perm)
answer: no it is not.
*therefore make an assertion about that*:
comb += Assert(self.ra[8:] == 0)
now let us move on to when any given "index" (in rs, not rb) is
greater than 64.
is it *ever* possible that perm[i] - which remember is the same
as ra[i] - that when "index" is >= 64, that perm[i]
(aka ra[i]) will be anything other than zero?
with m.If(idx < 64):
m.d.comb += perm[i].eq(rb64[idx])
answer: no, it cannot, because when index >= 64, perm[i] is never
set, and therefore will *always* be zero.
*therefore make an assertion about that*:
with m.If(index >= 64):
comb += Assert(ra[i] == 0)
and that just leaves the last one, which is annoyingly a little more
complicated, and best done probably by setting up that array rb64 again,
which will allow you to get at rb[index]
i "cheated" by doing a ridiculously-CPU-intensive cycling through
j=0..64 looking for a match against j == index.
--
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 @ 2020-05-19 0:14 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 #47 from Cole Poirier <colepoirier@gmail.com> ---
(In reply to Luke Kenneth Casson Leighton from comment #46)
> (In reply to Cole Poirier from comment #44)
>
> > 71 # 1. RA[8:64] should always equal 0
> > 72 # 2. If RB[i*8:i*8+8] >= 64 then RA[i] should equal 0
> > 73 # 3. If RB[i*8:i*8+8] < 64 then RA[i] should RS[index]
> >
> > For 2 and 3 should RB[i*8:i*8+8] actually be RS[i*8:i*8+8]?
>
> yes.
>
> > Or am I failing
> > to properly read/interpret the 'formalism'? In other words, I'm very
> > confused and looking for a way to better understand what I am doing, and
> > what I am supposed to be doing.
>
> you are expressing the expectations about the inputs and outputs of
> the module.
>
> these you express as assertions that must be true.
>
> (sby then checks those for you under *all* possible values of input and
> output).
>
> therefore you not only need to get the code right, you actually need to
> properly understand what it does... and express that in the assertions.
>
> example.
>
> from this line of code, is it *ever* possible that bits 8 and above
> will be *anything* other than zero?
>
> m.d.comb += self.ra[0:8].eq(perm)
>
> answer: no it is not.
>
> *therefore make an assertion about that*:
>
> comb += Assert(self.ra[8:] == 0)
>
> now let us move on to when any given "index" (in rs, not rb) is
> greater than 64.
>
> is it *ever* possible that perm[i] - which remember is the same
> as ra[i] - that when "index" is >= 64, that perm[i]
> (aka ra[i]) will be anything other than zero?
>
> with m.If(idx < 64):
> m.d.comb += perm[i].eq(rb64[idx])
>
> answer: no, it cannot, because when index >= 64, perm[i] is never
> set, and therefore will *always* be zero.
>
> *therefore make an assertion about that*:
>
> with m.If(index >= 64):
> comb += Assert(ra[i] == 0)
>
> and that just leaves the last one, which is annoyingly a little more
> complicated, and best done probably by setting up that array rb64 again,
> which will allow you to get at rb[index]
>
> i "cheated" by doing a ridiculously-CPU-intensive cycling through
> j=0..64 looking for a match against j == index.
Thank Luke, this is tremendously helpful. I think Michael explained it well,
but I was having a hard time fully comprehending because I'm still learning how
to read and write the 'language' of the cpu/rtl/circuit domain. Going to
re-read this a few times, then review nmigen tutorial by Robert Baruch, Claire
Wolf's 'Yosys Application Note 011: Interactive Design Investigation', so I
can better understand those 'languages'.
Tomorrow should I move on to trying to understand Wolf's verilog code for the
rvb equivalent of this operation? With regard to copyright, what should I do?
Am I supposed to write to her to ask for permission to use her code? Is that
even a possibility? I don't know how to proceed, what do you advise?
--
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 19may2020
From: Luke Kenneth Casson Leighton @ 2020-05-19 11:59 UTC (permalink / raw)
To: Libre-RISCV General Development
yesterday i went over the pipeline code, thought about register to function
unit allocation, wrote to Marketnext to alert their members to preinstall
the dev environment (thanks to Cole for helping with that), and recovered
slightly from late nights.
today i plan (started) drawing out the Function Units with their Read and
Write register port allocations.
the idea is to have everything on one page, from which it will be clear how
many (total) ports will be needed on each type of Regfile (INT, SPR,
Condition, Exception). then, the classes connecting FUs to Regfiles can be
designed.
except, i can't complete the diagram because i am blocked waiting for Jacob
to do the DIV and MUL pipelines, which we have found highlights clearly
what regs are needed in ways that a cursory examination does not reveal.
i will therefore try to do this task without having the pipeline code to
examine.
one of the primary purposes of the kan-ban daily updates is to identify
these kinds of blocks, when collaborating in a group.
we also have a few other blocks / task dependencies:
* Yehowshua, the pinouts need to be discussed with Tim as to what he would
like to see, before we put them into the pinmux. we cannot allocate that
to Rudi until it has been agreed by the team, and he has been waiting for
over 7 days (5 percent of the available time for completion of his task)
* Cole, the bperm proof (not exactly a block, more a confidence boost)
* Tobias, the countzero proof (not exactly a block, more a confidence boost)
* Alain, the kan-ban board, which will help us to identify task
dependencies a lot better.
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 323] New: create POWER9 MUL pipeline
From: bugzilla-daemon @ 2020-05-19 12:01 UTC (permalink / raw)
To: libre-riscv-dev
https://bugs.libre-soc.org/show_bug.cgi?id=323
Bug ID: 323
Summary: create POWER9 MUL pipeline
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: ---
a MUL pipeline is needed similar to the other pipelines in soc.fu, covering MUL
and MAC operations.
--
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 305] Create Pipelined ALU similar to alu_hier.py
From: bugzilla-daemon @ 2020-05-19 12:02 UTC (permalink / raw)
To: libre-riscv-dev
In-Reply-To: <bug-305-13@https.bugs.libre-soc.org/>
https://bugs.libre-soc.org/show_bug.cgi?id=305
Luke Kenneth Casson Leighton <lkcl@lkcl.net> changed:
What |Removed |Added
----------------------------------------------------------------------------
See Also| |https://bugs.libre-soc.org/
| |show_bug.cgi?id=323
--
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 323] create POWER9 MUL pipeline
From: bugzilla-daemon @ 2020-05-19 12:02 UTC (permalink / raw)
To: libre-riscv-dev
In-Reply-To: <bug-323-13@https.bugs.libre-soc.org/>
https://bugs.libre-soc.org/show_bug.cgi?id=323
Luke Kenneth Casson Leighton <lkcl@lkcl.net> changed:
What |Removed |Added
----------------------------------------------------------------------------
Assignee|lkcl@lkcl.net |programmerjake@gmail.com
See Also| |https://bugs.libre-soc.org/
| |show_bug.cgi?id=305
--
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 324] New: create POWER9 DIV pipeline
From: bugzilla-daemon @ 2020-05-19 12:03 UTC (permalink / raw)
To: libre-riscv-dev
https://bugs.libre-soc.org/show_bug.cgi?id=324
Bug ID: 324
Summary: create POWER9 DIV pipeline
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: ---
a DIV (and MOD) pipeline is needed similar to the other soc.fu pipelines.
--
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 324] create POWER9 DIV pipeline
From: bugzilla-daemon @ 2020-05-19 12:04 UTC (permalink / raw)
To: libre-riscv-dev
In-Reply-To: <bug-324-13@https.bugs.libre-soc.org/>
https://bugs.libre-soc.org/show_bug.cgi?id=324
Luke Kenneth Casson Leighton <lkcl@lkcl.net> changed:
What |Removed |Added
----------------------------------------------------------------------------
Assignee|lkcl@lkcl.net |programmerjake@gmail.com
See Also| |https://bugs.libre-soc.org/
| |show_bug.cgi?id=305
--
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 305] Create Pipelined ALU similar to alu_hier.py
From: bugzilla-daemon @ 2020-05-19 12:04 UTC (permalink / raw)
To: libre-riscv-dev
In-Reply-To: <bug-305-13@https.bugs.libre-soc.org/>
https://bugs.libre-soc.org/show_bug.cgi?id=305
Luke Kenneth Casson Leighton <lkcl@lkcl.net> changed:
What |Removed |Added
----------------------------------------------------------------------------
See Also| |https://bugs.libre-soc.org/
| |show_bug.cgi?id=324
--
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 324] create POWER9 DIV pipeline
From: bugzilla-daemon @ 2020-05-19 12:16 UTC (permalink / raw)
To: libre-riscv-dev
In-Reply-To: <bug-324-13@https.bugs.libre-soc.org/>
https://bugs.libre-soc.org/show_bug.cgi?id=324
--- Comment #1 from Luke Kenneth Casson Leighton <lkcl@lkcl.net> ---
register allocations, for fu.div.pipe_data.py
looks to be identical to ALUInputData and ALUOutputData.
https://libre-soc.org/openpower/isa/fixedarith/
# Divide
XO-Form divw* RT,RA,RB
Special Registers Altered:
* CR0 (bits 0:2 undefined in 64-bit mode) (if Rc=1)
* SO OV OV32 (if OE=1)
# Modulo
X-Form mod* RT,RA,RB
no special regs altered
--
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 323] create POWER9 MUL pipeline
From: bugzilla-daemon @ 2020-05-19 12:24 UTC (permalink / raw)
To: libre-riscv-dev
In-Reply-To: <bug-323-13@https.bugs.libre-soc.org/>
https://bugs.libre-soc.org/show_bug.cgi?id=323
--- Comment #1 from Luke Kenneth Casson Leighton <lkcl@lkcl.net> ---
there are actually two different types of MUL here.
* VA Form - 3 int in, no carry/overflow
* X Form - usual style just like ALU/Logical
my feelings are mixed as that is a lot of ports if they are combined. still,
actuslly, after some thought it is the same (after combining) port allocation
as Shift.
# Multiply-Add High Doubleword VA-Form
VA-Form
* maddhd RT,RA.RB,RC
prod[0:127] <- (RA) * (RB)
sum[0:127] <- prod + EXTS(RC)
RT <- sum[0:63]
Special Registers Altered:
None
--
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] New: create POWER9 TRAP pipeline
From: bugzilla-daemon @ 2020-05-19 12:28 UTC (permalink / raw)
To: libre-riscv-dev
https://bugs.libre-soc.org/show_bug.cgi?id=325
Bug ID: 325
Summary: create POWER9 TRAP pipeline
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: ---
we need a "trap" pipeline, or to decide if this is to be considered part of
Branch, due to the changes that trap makes to the PC. this by analysing the
register port allocation.
--
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 @ 2020-05-19 12:29 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
Luke Kenneth Casson Leighton <lkcl@lkcl.net> changed:
What |Removed |Added
----------------------------------------------------------------------------
See Also| |https://bugs.libre-soc.org/
| |show_bug.cgi?id=305
--
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 305] Create Pipelined ALU similar to alu_hier.py
From: bugzilla-daemon @ 2020-05-19 12:29 UTC (permalink / raw)
To: libre-riscv-dev
In-Reply-To: <bug-305-13@https.bugs.libre-soc.org/>
https://bugs.libre-soc.org/show_bug.cgi?id=305
Luke Kenneth Casson Leighton <lkcl@lkcl.net> changed:
What |Removed |Added
----------------------------------------------------------------------------
See Also| |https://bugs.libre-soc.org/
| |show_bug.cgi?id=325
--
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 @ 2020-05-19 12:30 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
Luke Kenneth Casson Leighton <lkcl@lkcl.net> changed:
What |Removed |Added
----------------------------------------------------------------------------
See Also| |https://bugs.libre-soc.org/
| |show_bug.cgi?id=313
--
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 313] Create Branch Pipeline for POWER9
From: bugzilla-daemon @ 2020-05-19 12:30 UTC (permalink / raw)
To: libre-riscv-dev
In-Reply-To: <bug-313-13@https.bugs.libre-soc.org/>
https://bugs.libre-soc.org/show_bug.cgi?id=313
Luke Kenneth Casson Leighton <lkcl@lkcl.net> changed:
What |Removed |Added
----------------------------------------------------------------------------
See Also| |https://bugs.libre-soc.org/
| |show_bug.cgi?id=325
--
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