* [libre-riscv-dev] [Bug 377] New: possible bug in Simulator Mem ld/st function
From: bugzilla-daemon @ 2020-06-12 14:58 UTC (permalink / raw)
To: libre-riscv-dev
https://bugs.libre-soc.org/show_bug.cgi?id=377
Bug ID: 377
Summary: possible bug in Simulator Mem ld/st function
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: ---
michael, hi,
it looks like there might be a bug in caller.py Mem ld/st. by accident
i asked a byte-store (stb) to write to address 0x9. the memory-dump
(fu/compunits/test/test_compunit.py) dumps out the internal dictionary
(sim.mem.mem.items()) however it uses sim.mem.ld to initialise the
nmigen Memory object.
* bytes per word is 8
* address 9 divided by 8 is 1
* remainder is also 1.
does this mean that the LD/ST is being word-order realigned?
Writing 0xee to ST 0x9 memaddr 0x1/1
width,rem,shift,mask 1 1 0x30 0xff
mem @ 0x1: 0xabeeef0187654321
None
carry already done? 0b0
get_cu_outputs 2 0
after got outputs, rd_rel, wr_rel, wrmask: 0b0 0b0 0b0
busy 1
busy 1
busy 1
busy 0
check cu outputs stb 3, 1(2) {}
check extra output 'stb 3, 1(2)' 0 0
sim mem dump
0 5432123412345678
1 abeeef0187654321
4 1828384822324252
nmigen mem dump
0 5432123412345678
1 abcdef018765ee21
2 0000000000000000
3 0000000000000000
4 1828384822324252
5 0000000000000000
--
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 12jun2020
From: Cole Poirier @ 2020-06-12 14:28 UTC (permalink / raw)
To: Libre-RISCV General Development
Yesterday: Frustration with gtkwave timiming diagrams...
Today: Frustration with my self for not thinking to zoom the gtkwave
timing diagrams out far enough to be able to see the changes in wire
state.... *bangs head on desk* Oh well, at least now I've figured it out
and can start making forward progress. Will spend today trying to
understand the relationship between a simple alu test in nmigen and the
corresponding timing diagram in gtkwave. Will also work on a blurb for
soliciting new members, and research some possible methods of
publicising our search for memebers.
Cole
_______________________________________________
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
* Re: [libre-riscv-dev] Fall 2022 Interfaces
From: Cole Poirier @ 2020-06-12 14:23 UTC (permalink / raw)
To: Libre-RISCV General Development; +Cc: libre-riscv-dev
In-Reply-To: <20200612140651.s724ovvshtjjziqo@topoi.pooq.com>
On Jun 12 2020, at 7:06 am, Hendrik Boom <hendrik@topoi.pooq.com> wrote:
> It's just possibe that we may not *need* 4k at 120 fps.
> Certainly there are many potential applications for our chip that
> don't
> need that kind of video.
I concur. If we are planning on selling 100 million units, shouldn't we
have several different levels of IO/GPU capablilty, each targeting
different power requirements? For example, a phone or a tablet doesn't
need 4k 120HZ video right?
Cole
_______________________________________________
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 375] Finding new project members who can help with complex detailed work such as LDSTCompUnit
From: bugzilla-daemon @ 2020-06-12 14:16 UTC (permalink / raw)
To: libre-riscv-dev
In-Reply-To: <bug-375-13@https.bugs.libre-soc.org/>
https://bugs.libre-soc.org/show_bug.cgi?id=375
--- Comment #2 from Cole Poirier <colepoirier@gmail.com> ---
(In reply to Luke Kenneth Casson Leighton from comment #1)
> (In reply to Cole Poirier from comment #0)
> > I think it would very helpful if we had more project members who can help
> > work on the really complex pieces of the HDL,
>
> yes. right now i am basically left doing absolutely everything, which is not
> to say that everyone else's contributions are not valuable, it's just that
> they're between 10 to 100 times less productive.
Fair enough, you have been optimizing your productivity for the past 25 years.
What's less fair is you not having enough time to do everything, so I'll make a
concerted effort starting today to try to find you more help.
> >I am willing to try and recruit more people. Do you think this is a good idea?
>
> if we want to hit the target, then yes.
>
> > If so, do you have
> > any suggestions as to how I should go about this?
>
> mailing lists, talk to people. there's funding available.
Ok, this gives me some good ideas. I'll write up a little blurb derived from
some of your past mails to other developer mailing lists. Then I'll try and
find hdl mailing lists, probably reaching out to university computer
science/electrical engineering graduate/undergraduate registrars? If anyone
else has ideas for mailing lists I should target of seek out that would be very
helpful.
--
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
* Re: [libre-riscv-dev] daily kan-ban update 11jun2020
From: Cole Poirier @ 2020-06-12 14:09 UTC (permalink / raw)
To: Libre-RISCV General Development; +Cc: libre-riscv-dev
In-Reply-To: <824f0b10-1853-eba7-2676-2984ebce9c38@gmail.com>
On Jun 11 2020, at 2:02 pm, Cesar Strauss <cestrauss@gmail.com> wrote:
> Interesting tools and readings, to share:
>
> * Pipelined Skid Buffer
>
> http://fpgacpu.ca/fpga/Pipeline_Skid_Buffer.html
[snip]
> * Understanding Latency Hiding on GPUs, by Vasily Volkov
>
> http://www2.eecs.berkeley.edu/Pubs/TechRpts/2016/EECS-2016-143.pdf
>
> Nice introductory chapter on internal GPU architecture.
Thanks so much for sharing these resources Cesar! I skimmed them
yesterday and I think they'll be really useful for me to understand some
of these complex computer architecture concepts.
Cole
_______________________________________________
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
* Re: [libre-riscv-dev] Fall 2022 Interfaces
From: Hendrik Boom @ 2020-06-12 14:06 UTC (permalink / raw)
To: libre-riscv-dev
In-Reply-To: <CAPweEDxSRNtU2Y8AS2Xc+mCP7wEpDtCUJkRuJ9k3bk5VHb9jKw@mail.gmail.com>
On Fri, Jun 12, 2020 at 02:30:42PM +0100, Luke Kenneth Casson Leighton wrote:
> On Fri, Jun 12, 2020 at 1:38 PM Yehowshua <yimmanuel3@gatech.edu> wrote:
> >
> >
> > > that may not be enough to handle framebuffers as well as run an OS.
> > > framebuffers are scarily demanding of memory bandwidth and they absolutely
> > > MUST have total absolute top priority.
> >
> >
> > The OMI interface has 3200MT/s which is the same spec as DDR4.
>
> 3200 million transactions/sec, let's assume 8 bytes. that's 3200
> megabytes per second. 25 gigabit / 8 equals 3.125 so that fits with
> the math , there. one 25 gigabit lane provides 3200 million 8-bit
> transactions per second (3.125).
>
> 4k is 1200 megabytes per second: 37.5% *percent* of the total memory
> bandwidth consumed by a 4k 60fps framebuffer! and if people wanted
> 120 fps instead, that's now a staggering 75% of a single OMI interface
> dedicated to the framebuffer!
>
> bear in mind that that is *just* the framebuffer hardware. the CPU -
> the GPU - is contending for the exact same memory resource.
>
It's just possibe that we may not *need* 4k at 120 fps.
Certainly there are many potential applications for our chip that don't
need that kind of video.
-- hendrik
_______________________________________________
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 376] Assess 40/45 nm 2022 target and interfaces
From: bugzilla-daemon @ 2020-06-12 13:52 UTC (permalink / raw)
To: libre-riscv-dev
In-Reply-To: <bug-376-13@https.bugs.libre-soc.org/>
https://bugs.libre-soc.org/show_bug.cgi?id=376
Yehowshua <yimmanuel3@gatech.edu> changed:
What |Removed |Added
----------------------------------------------------------------------------
CC| |yimmanuel3@gatech.edu
--- Comment #2 from Yehowshua <yimmanuel3@gatech.edu> ---
> 4k is 1200 megabytes per second: 37.5% *percent* of the total memory
> bandwidth consumed by a 4k 60fps framebuffer! and if people wanted
> 120 fps instead, that's now a staggering 75% of a single OMI interface
> dedicated to the framebuffer!
Raptor was saying that they were really looking for 1080p @60fps.
So we shouldn't be worried about 4k for now.
--
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 376] Assess 40/45 nm 2022 target and interfaces
From: bugzilla-daemon @ 2020-06-12 13:31 UTC (permalink / raw)
To: libre-riscv-dev
In-Reply-To: <bug-376-13@https.bugs.libre-soc.org/>
https://bugs.libre-soc.org/show_bug.cgi?id=376
--- Comment #1 from Luke Kenneth Casson Leighton <lkcl@lkcl.net> ---
from discussion at
http://lists.libre-riscv.org/pipermail/libre-riscv-dev/2020-June/008021.html
action points (edit and add as appropriate):
* to make a decision as to what top level of visual performance is needed
(maximum resolution, maximum framerate)
* calculate the framebuffer data transfer rate for that.
* to decide some top level of GFLOPs and MTriangles/sec (and other waffly
figures for the GPU side)
* estimate the memory bandwidth for the GPU side
* decide if we want to develop a hardware compression algorithm between memory
and framebuffer (an augmentation of Richard Herveille's RGBTTL framebuffer RTL)
* find out the power consumption in 45nm for these SERDES
* talk to Rudi to see if he can do a SERDES PHY for us
* talk to SymbioticEDA likewise
* talk to Dmitri from LIP6.fr to see what he can do, whether a stable 50 GHz
clock is achievable in 45nm (i understand from Dmitri that for a given target
clock - 25 gbit/sec in this case - you need the PLL to do double the clockrate)
--
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
* Re: [libre-riscv-dev] Fall 2022 Interfaces
From: Luke Kenneth Casson Leighton @ 2020-06-12 13:30 UTC (permalink / raw)
To: Libre-RISCV General Development
In-Reply-To: <8CDFD090-B84C-4013-BE76-D4815DDB7F5A@gatech.edu>
On Fri, Jun 12, 2020 at 1:38 PM Yehowshua <yimmanuel3@gatech.edu> wrote:
>
>
> > that may not be enough to handle framebuffers as well as run an OS.
> > framebuffers are scarily demanding of memory bandwidth and they absolutely
> > MUST have total absolute top priority.
>
>
> The OMI interface has 3200MT/s which is the same spec as DDR4.
3200 million transactions/sec, let's assume 8 bytes. that's 3200
megabytes per second. 25 gigabit / 8 equals 3.125 so that fits with
the math , there. one 25 gigabit lane provides 3200 million 8-bit
transactions per second (3.125).
4k is 1200 megabytes per second: 37.5% *percent* of the total memory
bandwidth consumed by a 4k 60fps framebuffer! and if people wanted
120 fps instead, that's now a staggering 75% of a single OMI interface
dedicated to the framebuffer!
bear in mind that that is *just* the framebuffer hardware. the CPU -
the GPU - is contending for the exact same memory resource.
we cannot possibly expect the GPU to operate effectively at that kind
of resource-starved level. bear in mind that not only the GPU 3D
Vulkan / bitmap / texturisation data has to go over that same
interface, the Operating System *also* has to swap in and out as well
to get that down to only 5% of total bandwidth, going to 64 bytes per
transaction would, if i am reading this correctly, require *eight* of
those 25 gigabit per second SERDES.
and that alone is going to be mental as far as power consumption is
concerned. i don't exactly know what the power consumption of a
single lane 25 gigabit SERDES is going to be, i am trying to recall
the conversations last year with Rudi when i visited him: it will
however be extremely important to find out.
i have a *vague* recollection of a number somewhere around 300mW. if
we assume that to be correct, then those 10x Rx and 14x Tx would be
somewhere around *eight watts* just for the SERDES alone!
so here we have the basis of actionpoints, which i would suggest be
put into the bugtracker (otherwise we'll lose track, which i'm
starting to do, already, thank you gmail)
* to make a decision as to what top level of visual performance is
needed (maximum resolution, maximum framerate)
* calculate the framebuffer data transfer rate for that.
* to decide some top level of GFLOPs and MTriangles/sec (and other
waffly figures for the GPU side)
* estimate the memory bandwidth for the GPU side
* decide if we want to develop a hardware compression algorithm
between memory and framebuffer (an augmentation of Richard Herveille's
RGBTTL framebuffer RTL)
* find out the power consumption in 45nm for these SERDES
* talk to Rudi to see if he can do a SERDES PHY for us
* talk to SymbioticEDA likewise
* talk to Dmitri from LIP6.fr to see what he can do, whether a stable
50 GHz clock is achievable in 45nm (i understand from Dmitri that for
a given target clock - 25 gbit/sec in this case - you need the PLL to
do double the clockrate)
l.
_______________________________________________
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 376] New: Assess 40/45 nm 2022 target and interfaces
From: bugzilla-daemon @ 2020-06-12 13:30 UTC (permalink / raw)
To: libre-riscv-dev
https://bugs.libre-soc.org/show_bug.cgi?id=376
Bug ID: 376
Summary: Assess 40/45 nm 2022 target and interfaces
Product: Libre-SOC's first SoC
Version: unspecified
Hardware: PC
OS: Mac OS
Status: CONFIRMED
Severity: enhancement
Priority: ---
Component: Milestones
Assignee: lkcl@lkcl.net
Reporter: lkcl@lkcl.net
CC: libre-riscv-dev@lists.libre-riscv.org
NLnet milestone: ---
we need to assess the interfaces and capabilities for the 2022 target ASIC
https://libre-soc.org/45nm_Fall2022/
--
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
* Re: [libre-riscv-dev] Fall 2022 Interfaces
From: Yehowshua @ 2020-06-12 12:38 UTC (permalink / raw)
To: Libre-RISCV General Development
In-Reply-To: <CAPweEDxge6pt4yx18nHY0Y6-f3rDOAF_Dq70YzbCqg8ud_DQCQ@mail.gmail.com>
> that may not be enough to handle framebuffers as well as run an OS.
> framebuffers are scarily demanding of memory bandwidth and they absolutely
> MUST have total absolute top priority.
The OMI interface has 3200MT/s which is the same spec as DDR4.
**Should be fine depending on line sized which is usually 8 - 64Bytes.
Yehowshua
_______________________________________________
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
* Re: [libre-riscv-dev] daily kan-ban update 11jun2020
From: Luke Kenneth Casson Leighton @ 2020-06-12 12:01 UTC (permalink / raw)
To: Libre-RISCV General Development
In-Reply-To: <824f0b10-1853-eba7-2676-2984ebce9c38@gmail.com>
On Thu, Jun 11, 2020 at 10:02 PM Cesar Strauss <cestrauss@gmail.com> wrote:
> Interesting tools and readings, to share:
i liked them, i added them to the "other" section on the resources page.
feel free to do that in future, it's what the Documentation Website
(wiki) is for.
https://libre-soc.org/resources/#index16h1
l.
_______________________________________________
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 374] add repo on github that points to git.libre-soc.org as a way to increase searchability
From: bugzilla-daemon @ 2020-06-12 11:55 UTC (permalink / raw)
To: libre-riscv-dev
In-Reply-To: <bug-374-13@https.bugs.libre-soc.org/>
https://bugs.libre-soc.org/show_bug.cgi?id=374
--- Comment #11 from Luke Kenneth Casson Leighton <lkcl@lkcl.net> ---
(In reply to Jacob Lifshay from comment #8)
> if someone wants to create the readme file, I can put it on my github
> account, which is probably the best one to use since I'm the one who's been
> part of libre-soc the longest who has a github account
good idea. we'll be able to do that when there is consensus.
--
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 374] add repo on github that points to git.libre-soc.org as a way to increase searchability
From: bugzilla-daemon @ 2020-06-12 11:55 UTC (permalink / raw)
To: libre-riscv-dev
In-Reply-To: <bug-374-13@https.bugs.libre-soc.org/>
https://bugs.libre-soc.org/show_bug.cgi?id=374
--- Comment #10 from Luke Kenneth Casson Leighton <lkcl@lkcl.net> ---
(In reply to Yehowshua from comment #7)
> > sadly whilst very "convenient", this indicates that our project endorses
> > github's
> > unethical practices. this includes censorship and government sanctioning.
>
> I disagree. Mirroring indicates that we do not endorse their practices.
it sounds reasonable, doesn't it? simply mirror the code, therefore
we're not actively endorsing unethical practices.
unfortunately, this is a false correlation, that takes some time to unpack.
normally i would explain it in detail, however unfortunately with whitequark
and many many other people who are prominent free software developers
with respected positions and responsibilities believing that i am ordering
people about, dictating to them what to think and what to do, i'm not going to
do that, because if i do so, it would seem like i am simply trying to force
my quotes beliefs quotes onto you.
so i'm going to leave this one to you to work through, to work out what the
false correlation is.
i will however give a starting point, with some questions, and a scenario.
this is something that i do all the time: i deploy reverse-engineering
skills - developed over a 25 year period - to *people*.
this skill is something that many people find alarming and scary (hence
the *real* reason why they think i am quotes ordering them about quotes).
imagine that there is a person, technically capable, who hears of libre-soc.
this imaginary person, they are both ambivalent and also completely unaware
of the ethical aspects of "Libre". nobody has ever had a conversation with
them about the unbelievably damaging aspects of arbitrary application of
technology without thought.
nobody has ever discussed the Hippocratic Oath as adapted to Technology
and Engineering with them.
(https://en.wikipedia.org/wiki/Order_of_the_Engineer#Oath)
now let us imagine that there are thousands, tens of thousands, hundreds
of thousands of such people. given that github probably has tens of millions
of users by now, this is not unreasonable.
however the reason for mentioning it is because that implies that there
are thousands to hundreds of thousands of engineers out there who will
see our project and have *NO IDEA* about the ethical responsibilities that
are associated with software and engineering in general.
thus if we have any way of influencing that, we achive a far larger impact
than we might otherwise have expected.
so that is background.
now let's ask some questions.
1. when these hypothetical (ethically ambivalent) people come to a straight
*MIRROR* of our code, on github, what will they see?
2. what *ACTION* will they take? in GUI web-development terms, what will
their browsing profile look like? what landing pages will they go to?
what pages will they click through?
in particular - bear in mind that you (and others) have been advocating that
all documentation - basically the entirety of the wiki - be moved to a "docs"
directory, because "this is how github does it". so this should be taken into
consideration as part of the answers to the questions.
now let us move on to what we would *like* their browsing profile to look
like:
3. where would we *LIKE* them to be browsing?
4. what pages and services would we *LIKE* them to be looking at?
so here is the key:
5. do we want these people to remain 100% on the github pages?
6. do we want them to be on http://libre-soc.org and http://systemes-libre.org?
so, let me know. (6) is really important. we might *actually* want people to
completely fail to come to our websites. we might actually want to send out
the message, "hey, we are completely technically incompetent, we are unable to
find the technical skill and take the responsibility for hosting our own
material, we completely abdicate that responsibility to an outside third party
known by the name 'github'."
whilst this may sound completely ridiculous and over-the-top, i *genuinely* and
honestly don't know, Yehowshua, because you keep advocating that we should use
these services.
so please note: there's no "demands" here. these are *open* questions that
people need to decide *for themselves* whether they should be discussed and
evaluated.
or ignored.
i have, personally, gone through these questions over a period of about 20
years and come to my own conclusions.
however it is clear to me - and continues to be clear - that many people simply
*do not* go rigorously through this process of evaluation.
i cannot force *anyone* to go through this process. that is not how it works.
--
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
* Re: [libre-riscv-dev] Kazan status
From: Luke Kenneth Casson Leighton @ 2020-06-12 11:27 UTC (permalink / raw)
To: Libre-RISCV General Development
In-Reply-To: <CALmHynRJ0xesoqqdsxN3jOcU-f8ehymb3M2+B0zmbkSCePE=7A@mail.gmail.com>
On Fri, Jun 12, 2020 at 8:12 AM Anton Bondarev
<anton.bondarev2310@gmail.com> wrote:
>
> Hi, all
> I'm Anton from Embox:)
hi anton :)
> I've some experience with etna-viv and mesa 3d (we ported etna-viv to
> Embox).
nice! this project? https://github.com/embox
> But I'm not sure that I'm an expert in GPU driver.
yes, porting of drivers is quite different from actually writing them.
at least, the main thing is: you actually know the codebase.
what did you actually do, what did the porting of etnaviv involve?
best,
l.
_______________________________________________
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
* Re: [libre-riscv-dev] Fall 2022 Interfaces
From: Luke Kenneth Casson Leighton @ 2020-06-12 11:21 UTC (permalink / raw)
To: Libre-RISCV General Development
In-Reply-To: <CAPweEDxge6pt4yx18nHY0Y6-f3rDOAF_Dq70YzbCqg8ud_DQCQ@mail.gmail.com>
(offlist conversation moved onlist after removing names).
phone messing formatting. blech. sorry.
the bandwidth requirements for video framebuffer are so insane i am leaning
towards suggesting use a multi scanline compression hardware algorithm.
similar to the old fax machines, huffman encoding, blah blah.
multi scanline because this will pick up vertical compression opportunities.
On Friday, June 12, 2020, Yehowshua <yimmanuel3@gatech.edu> wrote:
> two uses
>
1. Stand alone SBC
2. GPU for POWER systems, and any system really.
yep. this you can see on the original pages and descriptions i created.
glad to see you also worked it out.
> We can then literally provide identical GPU drivers to the host machine,
> except now, commands are sent over PCIE to LibreSOC where they are
> then issued as instructions on the hybrid CPU/GPU.
>
indeed. with associated insanity and complexity, for which an entire
special team will be needed.
And remember, LibreSOC already has video output -
>
only if Richard Herveille's RGBTTL interface is included.
so just send out the frame!
>
yyep :)
fortunately PCIe just memorymaps the framebuffer (i think)
so for nonaccelerated video it at least becomes braindead simple.
what is nice however is that by running an entire OS on the GPU, and it
being a full OS, a protocol can be invented which transfers video.
OR
just run the XServer *on the GPU*.
run xhost +
and... um... that's it.
done.
no need to write any drivers.
just treat the graphics card as if it was a networked Xserver.
> OK now to talk performance. When doing the discrete GPU, we can lift power
> restrictions.
> So lets say we double vector land widths, run the chip at 2GHz…
>
> We're talking 100 maybe 200GFLOPS.
>
yes.
it comes down to memory bandwidth at L0CacheBuffer.
4 core 800mhz dual issue is 1600 x 4 = 6.400 Giga vectors. and a FMAC is 2
already.
each vector is 4 FP32 therefore we are 50 GFLOPs
double the clock rate and 100 GFLOPs is achievable.
however to sustain that we need *SIXTEEN* LDST FunctionUnits.
and each one will have 2x LDST Ports.
32 ports @ around 160 bits wide
that's an internal routing of FIVE THOUSAND wires between the Function
Units and the L0 Cache Buffer.
in addition we will need a minimum of something mad like 8 way striped L1
data caches, and 16x 64 bit Wishbone Buses down into the L2 cache.
it's achievable, it's just going to be a handful.
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 318] fix LDSTCompUnit
From: bugzilla-daemon @ 2020-06-12 11:15 UTC (permalink / raw)
To: libre-riscv-dev
In-Reply-To: <bug-318-13@https.bugs.libre-soc.org/>
https://bugs.libre-soc.org/show_bug.cgi?id=318
--- Comment #28 from Cesar Strauss <cestrauss@gmail.com> ---
(In reply to Cesar Strauss from comment #26)
> I intend to do it in a general way, that will apply whenever you need to
> test a rel/go port.
>
> Basically, you tell the producer process what it needs to send, and it sends
> it whenever it receives a rel request, along with a go pulse.
>
> In the same way, you tell the consumer process what it should expect, and
> assert that it is presented by the port, along with the rel request, until
> the process emits the go pulse.
>
> In both cases, the rel->go delay will be parameterized, including a no-delay
> case.
Basically, what you already described in comment #11.
--
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 318] fix LDSTCompUnit
From: bugzilla-daemon @ 2020-06-12 11:13 UTC (permalink / raw)
To: libre-riscv-dev
In-Reply-To: <bug-318-13@https.bugs.libre-soc.org/>
https://bugs.libre-soc.org/show_bug.cgi?id=318
--- Comment #27 from Luke Kenneth Casson Leighton <lkcl@lkcl.net> ---
(In reply to Cesar Strauss from comment #26)
> In store():
>
> 538 if imm_ok:
> 539 yield dut.rd.go.eq(0b101)
> 540 else:
> 541 yield dut.rd.go.eq(0b111)
> 542 yield from wait_for(dut.rd.rel)
> 543 yield dut.rd.go.eq(0)
>
> Another case of multi-cycle assertion of go.
it works... it is however just a coincidence.
> This time, since there are multiple independent rel signals to wait for, a
> parallel test is the sensible way to go.
>
> I'll go ahead and implement it.
yes please. keep it as a separate test.
>
> I intend to do it in a general way, that will apply whenever you need to
> test a rel/go port.
oh good. it applies to ADR as well as ST so yes.
a base class that can be instantiated, this would be very handy.
> In both cases, the rel->go delay will be parameterized, including a no-delay
> case.
yes. like it.
--
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
* Re: [libre-riscv-dev] Fall 2022 Interfaces
From: Luke Kenneth Casson Leighton @ 2020-06-12 11:03 UTC (permalink / raw)
To: Libre-RISCV General Development
In-Reply-To: <CAC2bXD7fSit1+STs+WxMK4u2coY8GBnVgDw9mVjyq3VwUWdDDw@mail.gmail.com>
On Friday, June 12, 2020, Jacob Lifshay <programmerjake@gmail.com> wrote:
> On Thu, Jun 11, 2020, 19:59 Yehowshua <yimmanuel3@gatech.edu> wrote:
>
> > Was talking with Pine64 and Raptor about Fall2022 interfaces which I’m
> > trying to finalize.
> > We can do everything including HDMI, ethernet, DDR4(via OpenCAPI OMI),
> and
> > PCIE with SERDES lanes.
> >
>
> hdmi has issues with libre licensing and DRM content,
indeed. DVI however does not (and is pin compatible). HDMI is just a
protocol over DVI.
> if we're going that
> fast, we should just go with display port or similar, since it's an open
> standard, I think.
i added Richard Herveille's RGBTTL to the list. a TI TFP410a will convert
to at least 1600 x 1200 DVI no problem.
the TFP410a is a well known solution.
if we make it DDR RGBTTL then 1920x1080 is also no problem, via some
Chrontel ASICs.
>
>
> > We need 10rx and 14tx SERDES.
that is one hell of a lot.
> > 4rx and 4tx for the OMI will need to run at 25GHz for DDR4 over OMI.
that may not be enough to handle framebuffers as well as run an OS.
framebuffers are scarily demanding of memory bandwidth and they absolutely
MUST have total absolute top priority.
you cannot have a video scan line go blank. evverrr.
to give some idea: 1920 x 1080 is 2 million pixels. each pixel is 4
bytes. that is 64 million bytes per frame. 60 frames? that is 360
MEGABYTES per second.
DDR3 800 Mhz is only 6400 MB per second so we are looking at 20 percent of
memory bandwidth taken by feeding the framebuffer!!!!
if you want 4k that is now 1200 MEGABYTES per second just to feed the
frickin framebuffer!
doesn't even matter if it contains blank lines, that's 1.2 gigabytes per
second full of zeros!
> >
>
> 25GHz seems *really* fast, probably impractically so unless going to *lots*
> of effort, did you mean 2.5GHz?
no it really is 25GHz.
we will need Rudi from ASICS.ws for this. he can handle that kind of
work. when i went to visit him he was doing some SERDES PHY related work
at these speeds.
i do not know if 45nm can handle it, though.
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 318] fix LDSTCompUnit
From: bugzilla-daemon @ 2020-06-12 10:58 UTC (permalink / raw)
To: libre-riscv-dev
In-Reply-To: <bug-318-13@https.bugs.libre-soc.org/>
https://bugs.libre-soc.org/show_bug.cgi?id=318
--- Comment #26 from Cesar Strauss <cestrauss@gmail.com> ---
In store():
538 if imm_ok:
539 yield dut.rd.go.eq(0b101)
540 else:
541 yield dut.rd.go.eq(0b111)
542 yield from wait_for(dut.rd.rel)
543 yield dut.rd.go.eq(0)
Another case of multi-cycle assertion of go.
This time, since there are multiple independent rel signals to wait for, a
parallel test is the sensible way to go.
I'll go ahead and implement it.
I intend to do it in a general way, that will apply whenever you need to test a
rel/go port.
Basically, you tell the producer process what it needs to send, and it sends it
whenever it receives a rel request, along with a go pulse.
In the same way, you tell the consumer process what it should expect, and
assert that it is presented by the port, along with the rel request, until the
process emits the go pulse.
In both cases, the rel->go delay will be parameterized, including a no-delay
case.
--
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 375] Finding new project members who can help with complex detailed work such as LDSTCompUnit
From: bugzilla-daemon @ 2020-06-12 10:42 UTC (permalink / raw)
To: libre-riscv-dev
In-Reply-To: <bug-375-13@https.bugs.libre-soc.org/>
https://bugs.libre-soc.org/show_bug.cgi?id=375
--- Comment #1 from Luke Kenneth Casson Leighton <lkcl@lkcl.net> ---
(In reply to Cole Poirier from comment #0)
> I think it would very helpful if we had more project members who can help
> work on the really complex pieces of the HDL,
yes. right now i am basically left doing absolutely everything, which is not to
say that everyone else's contributions are not valuable, it's just that they're
between 10 to 100 times less productive.
> especially since Michael who
> has done a very significant amount of this work, now, and I think for the
> next few months(?) has less time to work on the project because he's writing
> his masters thesis. Since I am not presently capable of helping with most of
> the HDL code, especially the very complex pieces, I am willing to try and
> recruit more people. Do you think this is a good idea?
if we want to hit the target, then yes.
> If so, do you have
> any suggestions as to how I should go about this?
mailing lists, talk to people. there's funding available.
--
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
* Re: [libre-riscv-dev] Kazan status
From: Anton Bondarev @ 2020-06-12 7:11 UTC (permalink / raw)
To: Libre-RISCV General Development
In-Reply-To: <CAC2bXD6dumH+RkL9MCK-=aLukPp7z5ZmgVUxSapzaUT4hqVw3w@mail.gmail.com>
Hi, all
I'm Anton from Embox:)
I've some experience with etna-viv and mesa 3d (we ported etna-viv to
Embox). But I'm not sure that I'm an expert in GPU driver.
Anton
чт, 11 июн. 2020 г. в 23:56, Jacob Lifshay <programmerjake@gmail.com>:
> On Thu, Jun 11, 2020, 12:42 Luke Kenneth Casson Leighton <lkcl@lkcl.net>
> wrote:
>
> > On Thursday, June 11, 2020, Yehowshua <yimmanuel3@gatech.edu> wrote:
> >
> > > >
> > > > a full half of that first EUR 50,000 budget was allocated to
> > > > Kazan.
> > >
> > > On the note of Kazan, what is currently being done there?
> >
> >
> > nothing for the moment as we need all hands that have some nmigen
> > experience on the POWER9 core.
> >
>
> I'm taking a break on Kazan as stated above, Kazan's current status is that
> the SPIR-V to shader-compiler-ir translator and creating shader-compiler-ir
> is somewhere around half complete, I've been writing both of those together
> to allow rapid modifications to shader-compiler-ir to add support for
> SPIR-V features as I encounter them in the process of writing the
> translator. At that point, writing the shader-compiler-ir language
> reference docs will be the next step, since many of the invariants that the
> code follows is not really documented yet. Once shader-compiler-ir is
> stable and documented, it will be much more practical to have multiple
> people work on different parts of the compiler at the same time, since the
> basic interface will not change as much. After that, writing a reference
> interpreter will allow us to run some actual shaders, writing the SIMT to
> vector translator (the vectorization pass) will be enough to get to where
> we should be able to run shaders with the expected execution model and
> allow us to have all the parts for a working interpreted shader stack.
> Next, writing the shader-compiler-ir to LLVM IR translator and JIT
> compiler, creating a trait for a generic shader compiler backend (the part
> that translates the final shader-compiler-ir to machine code and allows
> executing the JIT compiled code or an interpreter) will allow us to have a
> MVP working shader compiler, as well as settling the interface between the
> rest of the Vulkan driver and the shader compiler crates.
>
> >
> >
> > > I contacted Anton from Embox, he worked on the Etna-viv drivers
> > > and he recently joined the mailing list.
> > >
> > > Anton, don’t be shy to say hi :)
> >
> >
> > hi anton, great to know you're around, we have a budget for 3D driver
> work
> > if you're interested?
> >
>
> Hi anton, glad to have additional people around who know a lot about 3D
> drivers.
>
> Jacob
> _______________________________________________
> libre-riscv-dev mailing list
> libre-riscv-dev@lists.libre-riscv.org
> http://lists.libre-riscv.org/mailman/listinfo/libre-riscv-dev
>
_______________________________________________
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
* Re: [libre-riscv-dev] Fall 2022 Interfaces
From: Yehowshua @ 2020-06-12 7:07 UTC (permalink / raw)
To: Libre-RISCV General Development
In-Reply-To: <CAC2bXD7fSit1+STs+WxMK4u2coY8GBnVgDw9mVjyq3VwUWdDDw@mail.gmail.com>
> On Jun 12, 2020, at 12:01 AM, Jacob Lifshay <programmerjake@gmail.com> wrote:
>
> 25GHz seems *really* fast, probably impractically so unless going to *lots*
> of effort, did you mean 2.5GHz?
I think its actually 5GHz - currently checking with somebody from IBM.
Yehowshua
_______________________________________________
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
* Re: [libre-riscv-dev] Kazan status
From: Lauri Kasanen @ 2020-06-12 5:52 UTC (permalink / raw)
To: libre-riscv-dev
In-Reply-To: <CAC2bXD6dumH+RkL9MCK-=aLukPp7z5ZmgVUxSapzaUT4hqVw3w@mail.gmail.com>
On Thu, 11 Jun 2020 13:56:22 -0700
Jacob Lifshay <programmerjake@gmail.com> wrote:
> I'm taking a break on Kazan as stated above, Kazan's current status is that
> the SPIR-V to shader-compiler-ir translator and creating shader-compiler-ir
> is somewhere around half complete,
Phoronix recently reported that the Vallium cpu implementation of
vulkan was already able to render Talos Principle. If work on it keeps
on at that pace, by Oct when Jacob is free again it will be much
further ahead.
Given this, it would make sense to consider moving to it over Kazan, or
at least checking the possibility.
- Lauri
_______________________________________________
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
* Re: [libre-riscv-dev] Fall 2022 Interfaces
From: Yehowshua @ 2020-06-12 5:46 UTC (permalink / raw)
To: Libre-RISCV General Development
In-Reply-To: <CAC2bXD7fSit1+STs+WxMK4u2coY8GBnVgDw9mVjyq3VwUWdDDw@mail.gmail.com>
>
> hdmi has issues with libre licensing and DRM content, if we're going that
> fast, we should just go with display port or similar, since it's an open
> standard, I think.
Perhaps - was reading Dave Shah’s convo on gitter…
He said Display Port is an absolute pain…
Yehowshua
_______________________________________________
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