<
kristianpaul>
haha
xiangfu has joined #m-labs
<
sb0>
got the misoc bios to compile with gcc 4.8.2 (and that patch) :)
<
sb0>
idk if it works
nicksydney has quit [Remote host closed the connection]
nicksydney has joined #m-labs
xiangfu has quit [Ping timeout: 240 seconds]
playthatbeat has joined #m-labs
mumptai has joined #m-labs
Alarm has joined #m-labs
mumptai has quit [Quit: Verlassend]
proppy has quit [Quit: Connection closed for inactivity]
xiangfu has joined #m-labs
xiangfu has quit [Ping timeout: 240 seconds]
nicksydney has quit [Remote host closed the connection]
nicksydney has joined #m-labs
Martoni has quit [Quit: ChatZilla 0.9.90.1 [Iceweasel 24.3.0/20140205221125]]
Martoni has joined #m-labs
Hawk777_ has joined #m-labs
Hawk777 has quit [*.net *.split]
bvernoux has joined #m-labs
<
bvernoux>
for those interested in a nice FPGA board
<
davidc__>
bvernoux: what USB3 phy/transceiver are you using?
<
bvernoux>
it is the only one existing TI USB1310A
<
bvernoux>
no other exist in fact ...
mumptai has joined #m-labs
<
davidc__>
bvernoux: gotcha, just a PHY not a txcvr
<
bvernoux>
only PHY
<
bvernoux>
all other stuff are done in FPGA
<
bvernoux>
with an USB 3.0 Device stack written in Verilog
<
bvernoux>
which should take less than 9Kcells
<
bvernoux>
anyway the USB 3.0 stack is not my work but Marshall Hecht's work
<
mumptai>
if there is a migen dev roadmap, is a vhdl export on it somewhere?
<
bvernoux>
Verilog is available here
<
bvernoux>
no there is not migen dev roadmap for it
<
bvernoux>
all is written from scratch in verilog
<
bvernoux>
and it is a huge work
<
mumptai>
i actually meant migen itself
<
bvernoux>
also USB 3.0 Device compliance test are pratically done
<
bvernoux>
missing feature are about power management
<
bvernoux>
and it will be only usb 3.0 device stack no host at all
<
mumptai>
but i can imagine that the protocol handling might be a bit tricky, does usb3 require downward compatibility of all all usb3 devices?
<
bvernoux>
yes it will support USB 2.0 too
<
bvernoux>
the USB 2.0 device was done before
<
bvernoux>
in respo there is both version
<
mumptai>
how fast can it actually go with a decent PC as host?
<
bvernoux>
IIRC it was > 200MB
<
bvernoux>
but was limited by the PC
<
bvernoux>
on FPGA side it shall go at maximum speed
<
bvernoux>
it is only limited by PC and USB 3.0 on host
<
bvernoux>
anyway the stack is not finished
<
mumptai>
and the fifo i/f is something like 32@100MHz ?
<
bvernoux>
phy pipe are 16bits
<
bvernoux>
16*250Mhz IIRC
<
bvernoux>
ij fact descrambling runs at only 125MHz
<
bvernoux>
but decoding 4 symbols at time
<
bvernoux>
so it work in parallell like if it will run at 500MHz
<
bvernoux>
which is very tricky on Cyclone4 ...
<
bvernoux>
4 symbols = 128bits
<
bvernoux>
128bits produce 32bits of data at end
<
bvernoux>
so 4Gigabits of data
<
bvernoux>
so a theorical maximum speed of 488MB/s
Alarm has quit [Quit: ChatZilla 0.9.90.1 [Firefox 27.0.1/20140212131424]]
proppy has joined #m-labs
bvernoux has quit [Quit: Leaving]
<
wpwrak>
sb0: relocated to .hk, closer to the cornucopia of gadgets ? :)
<
wpwrak>
wow. that came sudden. indefinite or just to visit for a while ?
<
wpwrak>
s/sudden/suddenly/ # headache vs. grammar, 1:0 :(
mumptai has quit [Quit: Verlassend]
Hawk777_ has quit [Ping timeout: 272 seconds]
Hawk777 has joined #m-labs