Untitled topic
1–10 of 11 posts
Re: undefined
#2"Flex-RV also integrates a programmable machine learning (ML) hardware accelerator inside the microprocessor and demonstrates new instructions to extend the RISC-V instruction set to run ML workloads. It is implemented, fabricated and demonstrated to operate at 60 kHz consuming less than 6 mW power."
Re: undefined
#3Re: undefined
#4Re: undefined
#5Re: undefined
#6But this is a 1960's era microprocessor: resistor-transistor logic running at 60 kHz. Something like the Apollo guidance computer, but at a 17 times slower clock speed.
The CPU is bit-serial: one bit at a time.
It saddens me they felt they needed to bolt-on some 4x8 bit multipliers. It feels like they just wanted to claim "the first flexible microprocessor with embedded ML capabilities".
https://en.wikipedia.org/wiki/Resistor%E2%80%93transistor_lo...
Re: undefined
#7I don't want to be dismissive or overly critical of new technology. But this is a 1960's era microprocessor: resistor-transistor logic running at 60 kHz. Something like the Apollo guidance computer, but at a 17 times slower clock speed. The CPU is bit-serial: one bit at a time. It saddens me they felt they needed to bolt-on some 4x8 bit multipliers. It feels like they just wanted to claim "the first flexible micropro…
SeRV at 60 kHz is going to do between 1000 and 2000 instructions per second, depending on the instruction mix.
As something to compare that against, an Apple ][ running AppleSoft BASIC runs `FOR I=0 TO N:T=T+I:NEXT I` at 250 loops per second.
On SeRV that loop is going to be something like ...
loop:
add a1,a1,a2
addi a2,a2,1
bge a0,a2,loop
... which will take 96 clock cycles per iteration, or about 600 loops per second at 60 kHz, so that's 2.4x faster than AppleSoft on an Apple ][. Which many people found fast enough to start a revolution.Yes, AppleSoft is using floating point numbers, which are slower than integers, but Integer BASIC only runs that same code at about 330 loops per second.
Original IBM Model 5050 PC runs the same loop at 415 loops per second in BASIC (380 for BASICA), if that's your reference ecosystem. Not a big difference. Still slower than SeRV at 60 KHz.
Re: undefined
#8I don't want to be dismissive or overly critical of new technology. But this is a 1960's era microprocessor: resistor-transistor logic running at 60 kHz. Something like the Apollo guidance computer, but at a 17 times slower clock speed. The CPU is bit-serial: one bit at a time. It saddens me they felt they needed to bolt-on some 4x8 bit multipliers. It feels like they just wanted to claim "the first flexible micropro…
Sometimes simply being programmable is enough, you don't need all that much speed. SeRV at 60 kHz is going to do between 1000 and 2000 instructions per second, depending on the instruction mix. As something to compare that against, an Apple ][ running AppleSoft BASIC runs `FOR I=0 TO N:T=T+I:NEXT I` at 250 loops per second. On SeRV that loop is going to be something like ... loop: add a1,a1,a2 addi a2,a2,1 bge a0,a2,…
Of course this doesn't alter the point about being as fast as the BASIC that early PCs came with, the only thing most end users had to write their own programs on them until things such as Turbo Pascal appeared (November 1983).
Re: undefined
#9Earlier quoted context omitted.
Sometimes simply being programmable is enough, you don't need all that much speed. SeRV at 60 kHz is going to do between 1000 and 2000 instructions per second, depending on the instruction mix. As something to compare that against, an Apple ][ running AppleSoft BASIC runs `FOR I=0 TO N:T=T+I:NEXT I` at 250 loops per second. On SeRV that loop is going to be something like ... loop: add a1,a1,a2 addi a2,a2,1 bge a0,a2,…
Oops, sorry, that RISC-V code will be 128 cycles per loop on SeRV not 96, as the branch needs 64 cycles. So about 470 loops per second not 600 on that flexible chip. Of course this doesn't alter the point about being as fast as the BASIC that early PCs came with, the only thing most end users had to write their own programs on them until things such as Turbo Pascal appeared (November 1983).
If you want an computer that you can wrap around a pencil, I bet it could be done simply using a very small conventional silicon chip embedded in a flexible medium. Perhaps some of this technology could help with interconnections.
Re: undefined
#10Earlier quoted context omitted.
Oops, sorry, that RISC-V code will be 128 cycles per loop on SeRV not 96, as the branch needs 64 cycles. So about 470 loops per second not 600 on that flexible chip. Of course this doesn't alter the point about being as fast as the BASIC that early PCs came with, the only thing most end users had to write their own programs on them until things such as Turbo Pascal appeared (November 1983).
One would have to run a similar BASIC interpreter on this processor to make that a reasonable comparison, and that would only highlight how ridiculously impractical this demo unit is. If you want an computer that you can wrap around a pencil, I bet it could be done simply using a very small conventional silicon chip embedded in a flexible medium. Perhaps some of this technology could help with interconnections.
Interpreted BASIC was the only thing most end users had on those computers, for quite a few years, until Turbo Pascal and Think Pascal/C changed everything.
Professional software was written in assembly language but at the start of the revolution most of the programs that people wrote themselves, or that you saw printed in magazines to type in yourself, were in BASIC.
It was fast enough for many tasks.