<azonenberg>
ZirconiumX: down the road i want to build a 2-way superscalar, hyperthreaded rv32i core to see how fast i can push it
<azonenberg>
in-order though, i'm not trying to be BOOM
<azonenberg>
i also want to try to build something tiny and size optimized a la picorv32, maybe even start with the picorv32 design, but hand tune it with RPMs to try and squeeze the max clock rate and least area i can out of it
<ZirconiumX>
azonenberg: I think there's potential in a very wide in-order core, akin to what Nvidia are trying with Carmel.
<azonenberg>
So, i have another design that isnt riscv based that i want to get back to
<ZirconiumX>
I presume you've heard of SERV
<azonenberg>
it's 3 IPC VLIW but all instructions are ternary and the pipeline is Y-shaped
<azonenberg>
specifically meant for cryptographic code
<azonenberg>
no integers, possibly not even arbitrary branching
<azonenberg>
it's meant to be a black box sitting between a tcp/ip stack and application layer, or similar
<azonenberg>
but trying to be agile enough that you can load new algorithms in the future rather than being tied to e.g. AES
<azonenberg>
and with guaranteed single cycle execution of all instruction, no caches, no timing side channels
<azonenberg>
dedicated rom/ram for instructions, memory mapped fifo for tx/rx data, constant storage/sbox rom, and scratchpad for temporary state
<azonenberg>
Basically there's three ALUs that each do r32 op r32 op imm32
<azonenberg>
except the third one operates on the output of the first two, instead of registers
<azonenberg>
so you only need one regfile write port despite doing 3 or 5 IPC (depending on if you count a ternary operation as one or two "instructions")
azonenberg_mobil has joined #yosys
<ZirconiumX>
CPU architecture is a very interesting thing
emeb_mac has joined #yosys
gsi__ has joined #yosys
vonnieda has joined #yosys
gsi_ has quit [Ping timeout: 272 seconds]
<azonenberg_mobil>
interesting indeed
pie_ has quit [Ping timeout: 245 seconds]
emeb_mac has quit [Ping timeout: 248 seconds]
emeb has left #yosys [#yosys]
PyroPeter has quit [Ping timeout: 252 seconds]
PyroPeter has joined #yosys
azonenberg_mobil has quit [Ping timeout: 252 seconds]
pie_ has joined #yosys
emeb_mac has joined #yosys
_whitelogger has joined #yosys
Max-P has quit [Ping timeout: 276 seconds]
Max-P has joined #yosys
citypw has quit [Ping timeout: 245 seconds]
ravenexp has joined #yosys
emeb_mac has quit [Ping timeout: 258 seconds]
citypw has joined #yosys
X-Scale has quit [Quit: HydraIRC -> http://www.hydrairc.com <- Nine out of ten l33t h4x0rz prefer it]
dys has quit [Ping timeout: 245 seconds]
gsi__ is now known as gsi_
dys has joined #yosys
dys has quit [Remote host closed the connection]
trmm has joined #yosys
<trmm>
back on the LPF versus parameters in the verilog; which one wins? if the LPF has a drive strength or slew rate, can the verilog override it?
citypw has quit [Ping timeout: 245 seconds]
ravenexp has quit [Quit: WeeChat 2.5]
Jybz has joined #yosys
Jybz has quit [Client Quit]
mirage335 has quit [Ping timeout: 252 seconds]
mirage335 has joined #yosys
pie__ has joined #yosys
pie__ has quit [Remote host closed the connection]
pie_ has quit [Read error: Connection reset by peer]
pie__ has joined #yosys
MoeIcenowy has quit [Quit: ZNC 1.6.5+deb1+deb9u1 - http://znc.in]
MoeIcenowy has joined #yosys
pie__ has quit [Remote host closed the connection]
pie__ has joined #yosys
pie__ has quit [Remote host closed the connection]
pie__ has joined #yosys
<ZirconiumX>
Right, I *believe* I have a live value table implementation
<ZirconiumX>
Or at least part of one
<ZirconiumX>
Is it preferable to use symbiyosys or yosys-smtbmc?
<ZirconiumX>
(maybe a question for ZipCPU)
rrika has quit [Ping timeout: 244 seconds]
<ZipCPU>
SymbiYosys
<ZipCPU>
It's much simpler than yosys-smtbmc, gives you access to more solvers, and handles formats and format conversions for you
rrika has joined #yosys
pie__ has quit [Remote host closed the connection]