lekernel changed the topic of #milkymist to: Mixxeo, Migen, Milkymist-ng & other Milkymist projects :: Logs: http://en.qi-hardware.com/mmlogs :: Mixxeo preorder lists.milkymist.org/pipermail/devel-milkymist.org/2013-May/003344.html
<azonenberg> Hawk777: is it non-copyleft open source?
<azonenberg> My entire SoC is BSD and i'm hoping to keep it that way
<azonenberg> My build system can optionally link with a GPL'd library and be dual-licensed
<azonenberg> But otherwise is pure BSD, all of the RTL is BSD too
zumbi has joined #milkymist
<Hawk777> azonenberg: not sure what GRLIB’s license is; you should check it yourself. I don’t need to know as I’m only using it in stuff that will probably never be released anyway.
<azonenberg> I see
<azonenberg> I'll look into it
<azonenberg> I need to learn more hard number-theory math anyway
<azonenberg> and opencores has a complete lack of elliptic curve stuff last i looked
<azonenberg> even the RSA modules they had were horrible
<larsc> I think gems are rare on opencores in general
kilae has joined #milkymist
<azonenberg> Lol, true
<azonenberg> the last CPU i tried to use from there didn't even synthesize on the platform they said it was designed for :p
bhamilton has joined #milkymist
bhamilton has quit [Client Quit]
bhamilton has joined #milkymist
lekernel has joined #milkymist
bhamilton has quit [Client Quit]
lekernel has quit [Ping timeout: 260 seconds]
lekernel has joined #milkymist
antgreen has quit [Read error: Operation timed out]
antgreen has joined #milkymist
antgreen has quit [Ping timeout: 246 seconds]
mumptai has joined #milkymist
Hawk777 has quit [Ping timeout: 260 seconds]
Hawk777 has joined #milkymist
antgreen has joined #milkymist
<lekernel> read: 5075Mbps write: 5195Mbps all:10270Mbps
<lekernel> new memory controller on the M1 is starting to do quite good :)
<larsc> what were the previous numbers?
stekern has quit [Read error: Operation timed out]
stekern has joined #milkymist
Scopeuk-AFK is now known as Scopeuk
<mumptai> is that for a 16bit 333MHz DDR interface?
<lekernel> 32-bit but yes
<lekernel> (166MHz clock)
<mumptai> wow, that pretty tight pipelining then ;)
<mumptai> is it using the mcb, or something homegrown?
<lekernel> no xilinx IP whatsoever except the IOSERDES
<ysionneau> no mcb
<lekernel> all the rest is open source migen design
<lekernel> I'm not even sure the MCB is that fast :)
<mumptai> ioserdes was the 1/4 rate thingy?
<ysionneau> 17:17 < lekernel> read: 5075Mbps write: 5195Mbps all:10270Mbps < you are doing read and write "at the same time"?
<ysionneau> I mean those numbers are obtained by doing read *and* write?
<ysionneau> or just read and then just write
<lekernel> no. but that controller would do some transaction reordering to group read and writes together to minimize bus turnaround times.
<ysionneau> great!
<azonenberg> lekernel: the IOSERDES are not IP
<azonenberg> they're hard peripherals
<lekernel> yeah, I know, hard IP if you prefer
<lekernel> and you could do that with the fabric too - it's even faster xD http://hamsterworks.co.nz/mediawiki/index.php/Spartan_6_1080p
<azonenberg> lolwut?
<azonenberg> It's not faster
<azonenberg> its a hackjob
<azonenberg> if you read, the PLL jitter prevents it from being fully standard compliant
<lekernel> yeah, don't get me started on slowtan6 PLL jitter. but that's a different problem, which you could workaround with an external PLL chip.
<azonenberg> Lol
<azonenberg> Or use 7 series
<azonenberg> Have you used 7 yet?
<lekernel> no
<lekernel> except running lm32 synthesis on vivado, which looked quite good. of course, maybe the bitstream it produced in so little time would not work when loaded ;)
<azonenberg> lol
<lekernel> ysionneau, goes down to 9169Mbps with simultaneous read/write
<ysionneau> still nice
<lekernel> could perhaps be optimized in a simple way by changing r/w switch decision thresholds, haven't tried
<ysionneau> what do you mean by "all:10270Mbps" ?
<lekernel> total data rate going into + out of the DRAM chips is 10270Mbps
<ysionneau> because you're doing read and write at the same time during the same test
<lekernel> I have a write core and a read core
<ysionneau> are the results good for Mixxeo? for dual dvi input + vga output and maybe previsualization on two small LED screens ?
<lekernel> I can run them simultaneously or not
<ysionneau> ok understood
<lekernel> it'll be top-notch AMOLED :) finally got the german bureaucrats to ship them, yay!
<lekernel> getting the IO specification from I€€€, on the other hand, is another problem
<lekernel> worst case it should not be too hard to RE from the various bits of info they let slip
<ysionneau> :)
<ysionneau> so you think there will be enough ram bandwidth for all of this
<ysionneau> maybe trying to reduce the ram used by lm32 can save some bandwidth, by using small sleep loops entirely contained in the cache
<ysionneau> and just serve interrupts
<lekernel> yes, 720p will use 7G
<lekernel> lm32 is <200M even with a lot of cache misses
<lekernel> also, mixxeo will have 20% faster memory (DDR400 instead of DDR333)
<lekernel> and a -3 slowtan6 that should run at 100MHz almost without excessive headache
<ysionneau> so you will be able to access the DDR400 with a 200 MHz clock?
<lekernel> yes
<lekernel> hopefully
<ysionneau> that would be cool!
<GitHub51> [milkymist-ng] sbourdeauducq pushed 2 new commits to master: http://git.io/wGKJiQ
<GitHub51> milkymist-ng/master 2f662bf Sebastien Bourdeauducq: software/videomixer: fix overflow in memory bandwidth computation
<GitHub51> milkymist-ng/master f3e07db Sebastien Bourdeauducq: software/memtest: basic test
<lekernel> one of the most satisfying commit messages ever
<ysionneau> very tricky lurking bug?
<lekernel> overflow in memory bandwidth computation ;)
<ysionneau> ahah it was showing a very small number?
<ysionneau> "why is my controller so slow?"
<lekernel> it was showing garbage above a certain bandwidth :)
kilae has quit [Quit: ChatZilla 0.9.90 [Firefox 22.0/20130618035212]]
<GitHub100> [linux-milkymist] larsclausen pushed 1 new commit to master: http://git.io/gs12YQ
<GitHub100> linux-milkymist/master 4702860 Lars-Peter Clausen: lm32: Fix milkymist_timer_init prototype...
Scopeuk is now known as Scopeuk-AFK
antgreen has quit [Read error: Operation timed out]
<GitHub46> [milkymist-ng] sbourdeauducq pushed 1 new commit to master: http://git.io/3_N65g
<GitHub46> milkymist-ng/master d753c52 Sebastien Bourdeauducq: lasmicon: fix FSM reset state with delayed_enter
<kristianpaul> could the mcb do it that fast?
<kristianpaul> or that write/read troughput is getting closer to the built-in hw?
lekernel has quit [Quit: Leaving]
mumptai has quit [Ping timeout: 240 seconds]