futarisIRCcloud has joined ##openfpga
_whitelogger has joined ##openfpga
Degi has quit [Ping timeout: 258 seconds]
Degi has joined ##openfpga
genii has quit [Quit: See you soon.]
kiboneu_ has quit [Quit: bye]
kiboneu has joined ##openfpga
kiboneu has quit [Quit: bye]
kiboneu has joined ##openfpga
Bike has quit [Quit: leaving]
mumptai_ has joined ##openfpga
mumptai has quit [Ping timeout: 260 seconds]
<TD-Linux> yosys + vpr + fasm
<whitequark> so, why doesn't quicklogic use nextpnr?
<whitequark> the feedback i've seen about vpr is that it is "more of a pain to use than proprietary toolchains" so i'm not looking forward to use it
<TD-Linux> slightly annoying their devkits aren't available through normal distributors
<daveshah> CC or quicklogic?
<TD-Linux> quicklogic
<daveshah> I think they are planning to sort out something like that soon
<TD-Linux> at least they have a working online shop it seems
<tnt> They're also working on a bit more convenient package than 0.4mm bga or 0.35mm wlcsp ...
<TD-Linux> the EOS-S3 looks like a qfn?
<tnt> The QFN package was discontinued at their fab so it's not available to buy, it's only used on the quickfeather.
<TD-Linux> bleh
<TD-Linux> the ice40 tqfp100 package is so good
<tnt> iew ... tqfp ...
<TD-Linux> it's enormous but 2 layer boards!
<tnt> 4l is cheap nowadays, no reason to go for 2l.
<tnt> also if you _really_ insist you can make a 2l with QFN.
<TD-Linux> for the quicklogic in particular I kind of wanted to use it with power electronics which push the board a lot bigger. though there is always the multiple PCBs option.
OmniMancer1 has joined ##openfpga
OmniMancer has quit [Ping timeout: 264 seconds]
OmniMancer has joined ##openfpga
OmniMancer1 has quit [Ping timeout: 260 seconds]
Lord_Nightmare has quit [Quit: ZNC - http://znc.in]
Lord_Nightmare has joined ##openfpga
<Lofty> whitequark: care to design my nextpnr databases for me? /s
<whitequark> i mean
<whitequark> maybe? i'm not actually sure what's involved but i'm not very happy with the database size either way
<TD-Linux> pretty sure my cpu cache is equally unhappy
<daveshah> yes, the current approach is not great
Richard_Simmons has joined ##openfpga
Richard_Simmons2 has quit [Ping timeout: 260 seconds]
Asu has joined ##openfpga
mumptai_ has quit [Quit: Verlassend]
mumptai has joined ##openfpga
X-Scale` has joined ##openfpga
X-Scale has quit [Ping timeout: 258 seconds]
X-Scale` is now known as X-Scale
Bike has joined ##openfpga
mossmann has quit [Ping timeout: 265 seconds]
mossmann has joined ##openfpga
genii has joined ##openfpga
jfcaron has joined ##openfpga
hackkitten has quit [Ping timeout: 272 seconds]
hackkitten has joined ##openfpga
craigjb_ has joined ##openfpga
craigjb has quit [Ping timeout: 264 seconds]
hackkitten has quit [Ping timeout: 246 seconds]
hackkitten has joined ##openfpga
<steve|m> anyone knows what FPGA this could be? looks like lattice to me: https://aliexpress.com/item/4000707253428.html or https://aliexpress.com/item/33059305651.html
hackkitten has quit [Ping timeout: 265 seconds]
<daveshah> based on the package, I'd say maybe ECP3
hackkitten has joined ##openfpga
<cr1901_modern> Is there an ECP4 or ECP2, or is this another case where Lattice can't count?
<Hoernchen> lol
<steve|m> which reminds me of https://youtu.be/VCK9l9gmyDk
<Lofty> cr1901_modern: https://www.eejournal.com/article/20111129-lattice/ <-- apparently so?
<zyp> as far as I can tell, theres ECP, ECP2, ECP3 and ECP5, so I guess they only skipped ECP4
<gruetzkopf> ecp4 nearly existed
<gruetzkopf> they were definitly planning on it
<cr1901_modern> Just saying... wouldn't have been surprised. B/c Lattice really can't count :P.
<Hoernchen> there are quite a few weird devices hidden in diamond, or so i heard..
<cr1901_modern> "LCXMO2-10000" or something like that is one of them
<Hoernchen> that is the one i was thinking of
<daveshah> Not Diamond, but the iCEcube sim models have some interesting primitives for the cancelled iCE40MX series
<daveshah> and some early iCE40 brochures mention a HX16K
<gruetzkopf> the ecp4-190 woul've been fun
<Hoernchen> the bar to look at icecube was always too low for me
<gruetzkopf> 12* 6G serdes
<daveshah> I think Lattice cancelled these as there was too much overlap
<daveshah> with their existing lines
<cr1901_modern> I like how the "8k" part has 6 times the number of LUTs of the "1k" part
<daveshah> It's not as bad as the CrossLink where the "40" part is closer to 30k LUTs
<daveshah> CrossLink NX
<cr1901_modern> 8KiLALUTs is 7680 LUTs (Kilo LAttice LUTs)
<daveshah> lol
<TD-Linux> with 3 points I can extrapolate this quadratically
<hitomi2500> I remember reading a sad story about ECP4. The one about cancelling mass production a few months after the samples were shipped to customer. Could be hoax though.
<daveshah> Yes, they definitely shipped some samples
<Hoernchen> and then the customer said ni?
<daveshah> I don't exactly know why it was cancelled, but I think it was mostly Lattice not being able to compete with Intera/Xilinx in the big FPGA space
<Lofty> Wasn't their goal to save money by using 65nm?
<daveshah> I think ECP4 was 40nm like ECP5
<hitomi2500> Yeah, that's probable pretty much the reason they did cancel the series. The ECP5 was launched later with a message like "small yet powerful"
<daveshah> oh no, looks like ECP4 was indeed going to be 65nm
OmniMancer has quit [Quit: Leaving.]
emeb has joined ##openfpga
indy has quit [Ping timeout: 256 seconds]
mkru has joined ##openfpga
mkru has quit [Quit: Leaving]
<awygle> where's my ecp6 :(
<tnt> ecp-nx you mean
<Lofty> Probably one of the new FDSOI chips
<daveshah> Yeah ECP NX it will almost certainly be, don't have any official confirmation
<daveshah> All Lattice have said is that there are two new families on the FDSOI process coming this year
<tnt> probably on the 31st december at 23:59 ...
<daveshah> My guess is that would be an iCE40 successor and an ECP5 successor, but it is also possible they will do a "security" device instead of the iCE40, xo3d esque
<daveshah> yeah lol
<awygle> i want an ecp* with 10 Gbps transceivers
<daveshah> one was supposed to be in H1, so they have two weeks
<awygle> please and thank you
<tnt> Well .. 12.5G really ...
<daveshah> not happening, I saw a primitive name somewhere in radiant of "PCS6" so I would assume 6 or 6.25G
<awygle> either that or one that's cheaper than the current ones
<daveshah> not happening either, the NX series so far are much more expensive...
<daveshah> I don't think you can really get cheaper than ECP5 tbh
<tnt> 6.25G would be nice at least you can do 10G with just 2 of them, that's an improvement.
<tnt> 10G ethernet I mean.
<gruetzkopf> ECP4 had 6G apparently
<tnt> Probably why it doesn't exist ... they probably tried setting Vcore=10V to reach that and realized it wasn't lasting long enough.
<tnt> I'm kind of curious is 6.25G would work on a ecp5-5g ...
<daveshah> I wouldn't be too surprised if it works, but performance could equally be pretty bad given 5G is arguably marginal already
<awygle> you can only do ethernet with 2x6.25G with rxaui which is mostly not available :(
<daveshah> Personally it's 8x SERDES that would be nice
<awygle> yeah 8x serdes would also be good. or 16x
<awygle> tbf tho dave, the NX series so far has a bunch of weird custom IP
<daveshah> yes
<daveshah> but that's partly because CrossLink == weird custom IP
<awygle> not necessarily indicative of fab process impact on prices
<daveshah> it's mostly not custom either
<tnt> awygle: I found a microsemi phy which is (1) at digikey and (2) has register level doc public ...
<daveshah> the PCIe is from NorthWest logic, the MIPI from mixel
<awygle> tnt: without NDA/sign-in?
<tnt> yeah.
<awygle> i would be interested in that if you have a link
<gruetzkopf> what's the other end? -R ethernet?
<gruetzkopf> oh wow, even has a STS-192c bitrate mode
<awygle> huh weird, it says "registered user" on the microsemi website
<awygle> what about the VSC8486?
<tnt> Don't think it does 2*6.25G
<tnt> why did you want that one ?
<awygle> it does XGMII
<tnt> yeah, that's a lot of pins :p
<awygle> yup :p but you don't need serdeses!
<gruetzkopf> step2: four of them and a lot of byte shifting for 40GE without serdes
<awygle> now that's a _lot_ of pins
<gruetzkopf> 264 iirc?
<tnt> With a 756 caBGA ECP5 you might have enough ...
<tnt> 365 IOs.
<gruetzkopf> that's solidly into stupid tech demo category
<tnt> 8 * 32bit at 156.25 MHz DDR busses ... yeah ....
<gruetzkopf> hm, i need to check which stripe size 40GE uses
<gruetzkopf> "Data is striped across the electrical lanes 66 bit blocks at a time (round robin)"
<gruetzkopf> okay, not too annoying..
<Hoernchen> lol
<gruetzkopf> XLMII and CGMII were proposed
<gruetzkopf> with the latter using a 320bit wide (per direction) interface for 100GE
<gruetzkopf> the lane align words are a problem though
_whitelogger has joined ##openfpga
sgstair_ has joined ##openfpga
X-Scale` has joined ##openfpga
kc8apf_ has joined ##openfpga
laintwo has joined ##openfpga
mossmann_ has joined ##openfpga
ZirconiumX has joined ##openfpga
englishm1 has joined ##openfpga
keesj_ has joined ##openfpga
eightdot_ has joined ##openfpga
sensille_ has joined ##openfpga
mossmann has quit [*.net *.split]
X-Scale has quit [*.net *.split]
Lofty has quit [*.net *.split]
sgstair has quit [*.net *.split]
kc8apf has quit [*.net *.split]
jeandet has quit [*.net *.split]
qu1j0t3 has quit [*.net *.split]
sensille has quit [*.net *.split]
laintoo has quit [*.net *.split]
Ekho has quit [*.net *.split]
englishman has quit [*.net *.split]
keesj has quit [*.net *.split]
eightdot has quit [*.net *.split]
X-Scale` is now known as X-Scale
kc8apf_ is now known as kc8apf
ZirconiumX is now known as Lofty
jeandet has joined ##openfpga
qu1j0t3 has joined ##openfpga
Ekho has joined ##openfpga
sensille_ is now known as sensille
esden has quit [Read error: Connection reset by peer]
elms has quit [Ping timeout: 260 seconds]
eddyb[legacy] has quit [Ping timeout: 244 seconds]
esden has joined ##openfpga
edmoore has quit [Read error: Connection reset by peer]
ovf has quit [Ping timeout: 246 seconds]
edmoore has joined ##openfpga
esden has quit [Ping timeout: 256 seconds]
edmoore has quit [Ping timeout: 256 seconds]
edmoore has joined ##openfpga
edmoore has quit [Read error: Connection reset by peer]
Asu has quit [Remote host closed the connection]
Asu has joined ##openfpga
sgstair_ is now known as sgstair
OmniMancer has joined ##openfpga
Asuu has joined ##openfpga
esden has joined ##openfpga
Asu has quit [Ping timeout: 264 seconds]
edmoore has joined ##openfpga
ovf has joined ##openfpga
Asuu has quit [Quit: Konversation terminated!]
elms has joined ##openfpga
eddyb[legacy] has joined ##openfpga
jfcaron has quit [Ping timeout: 256 seconds]
mumptai has quit [Remote host closed the connection]
emeb has quit [Quit: Leaving.]
emeb_mac has joined ##openfpga
OmniMancer1 has joined ##openfpga
OmniMancer has quit [Ping timeout: 264 seconds]