Thoughts?

NVIDIA throwing stacks of money at it
Western Digital throwing stacks of money at it
chinks throwing money at it to avoid ARM royalties
pajeets throwing money at it because MUH NATIONAL CPU
THALES throwing money at it because MUH MILITARY-INDUSTRIAL COMPLEX
universities around the world shitting out EE/CE graduates trained on it
buckle up son, the future is RISC-V

I need a CPU that is not backdoored asap. Should I early-buy this shit?

Dead in the water now that the MIPS ISA is getting opened up. Suck on my SIMD, RISC-V nigglets.

No, the RISC-V tinkerboards are gay and overpriced. Get a POWER9 Talos II workstation instead.

Attached: 027.png (1000x1000, 137K)

>Talos II
>$2200 for entry level
>company based in IL
>IBM-made CPU
>'uncucked'

Attached: 1533073087878.png (853x482, 433K)

x86 is industry standard. there is no good reason to use this shit.

youtube.com/watch?v=IfHG7bj-CEI

>domestically made hardware that's powerful and supported by a huge company
Are you fucking retarded?

Attached: 1463921189168037824.jpg (235x225, 25K)

it's a hardware version of java

this. It's legit.

The ISA is pretty great, and I expect it will see adoption pretty soon in things such as the kind of microcontrollers where ARM currently dominates, where the ARM licensing fees actually matter. From there, I wouldn't be surprised if it works its way up to SoCs. Probably hobby-class devices like RPis at first, but it may well spread from there, especially when people used to the hobby devices want to start using the professionally.

What's arguably far more interesting, however, is the vector extension, which looks to turn out truly great and be far superior to AVX or ARM's SVE, and I expect will actually make enough of a practical difference that such things as HPC and professional rendering will want it, and it may very well work its way down from there into workstations for the same reasons.

If the latter scenario turns out true, then that will drive high-end implementations, and having both the low-end of the market in µCs and SoCs, and the high-end of the market, it might eventually make some sort of impact on personal computers as well. Compatibility will of course be king for the foreseeable future, but perhaps some sort of Talos system will be released for those who want it.

MIPS is no competitor to RISC-V if you're starting pretty much from scratch implementation-wise anyway. RISC-V is much better thought out in terms of available instructions and encoding, and doesn't have MIPS' legacy crap in delay slots and all that which hasn't been relevant for 30 years.

wat

>As an instruction set, I like ARM better, the way 32 bits per instruction are utilized to it's full potential is so comfy.
The compressed instruction set allows you to save a lot more than that, though.

I want to read the Computer Organization and Design, The Hardware/Software Interface soon, is the RISC-V version recommended?

Attached: CODRVcover-242x300.jpg (242x300, 23K)

can it run firefox?

content.riscv.org/wp-content/uploads/2018/07/DAC-2018-Western-Digital-Zvonimir-Bandic.pdf

Everything at work is still ARM and 8051 (t. embedded dev), but I could see RISC-V becoming wildly popular in the next 5-10 years.

yes

I’d just stick with mips. Also that book was way too verbose for me.

Attached: gzql4cqeng301.jpg (645x729, 40K)

Wrong.
t. verification engineer at risc-v startup having so much shit to do that two lifetimes wouldn't be enough, thanks for going bonkers you filthy yellow niggers

Its practically identical. There are a few passages that address problems encountered in other ISAs and what RISCV opted to do to prevent them, but otherwise they took the same content and translated it to RISCV

Yes, it can.

"Too verbose" and "too hard to understand" are very different things, and if anything negatively correlated. Your picture represents yourself well.

it is good then

>verification engineer

Attached: interaction_engineer.jpg (687x389, 38K)

Not him, but you have absolutely no idea what you're talking about.

RISC is good.

Attached: DifaA27UEAAovIn.jpg (1200x515, 66K)

t. him

not him nor him as well, but wtf verification engineer is?

Verification in semiconductor manufacturing is basically what you'd call testing in software development. It's about making sure that manufactured chips work as specified and intended, and it is not a trivial task at all. I don't work in the field myself, but I've read in several articles and research papers that verification has risen to be among the most challenging areas altogether as chip complexity increases.

There is a widespread rumor that Intel's recent security problems is a consequence of them cheapening out on verification in order to become more """agile""".

Oh yeah we do that as well, but some other dedicated department in different building does that. Allegedly they discovered shitton of hw and calibration mistakes in sensors from our vendors (and probably in our device manufacturing as well).

I want to stick my dick in maki

you didn't know FPGA???
retard as fuck

>There is a widespread rumor that Intel's recent security problems is a consequence of them cheapening out on verification in order to become more """agile""".
That is not a verification problem, but a design problem. And paid well for some time. When you are in bed with the big shots you can get away with it.

Also an embedded dev (automotive stuff) here, I've seen way more AVRs than 8051s in daily usage. What kind of market do you work on?

get with the times gramps, ARM is already taking over

This is either the dumbest thing I've read on this board, or the best troll.

What do you think is running on your mobile phone, your router, your fridge, your sous vide machine, your VoIP phone at work, and your tablet?

the future lies in reversible computers and software.

Ah, sorry, I didn't actually refer to the Meltdown and following fiasco, but to some of the earlier problems. I don't recall them all, but one was the hyperthreading lock-up bug, which would have been a verification problem more than a design problem.

Reversible computing does not solve a problem that we will be having in many, many generations yet. It isn't even clear if we'll ever get to the point where that's a problem.

The dumb thing is when shills disingenuously pretend anyone on Jow Forums cares about iToys and the Internet of Shit. Personal computing is always the default context, and in that context he's still pretty much correct. Nobody gives a shit if their next cloud enabled prostate massager is going to be built around a RISC-V or ARM microcontroller.

To be sure, if there were RISC-V processors for PCs that were at least within a factor of two or three within the performance of Intel or AMD, I'd switch in a heartbeat.

But there never will be, and even if there is something remotely competitive it will undoubtedly be so expensive, niche and hard to get a hold of that you will gawk at it just like you gawked at Alpha, MIPS and PowerPC systems in the '90s that actually had tangible advantages to offer consumers that the comically oversized SBCs and "secure workstations" of today just don't over commodity hardware.

This we will likely continue to laugh at "year of the RISC desktop" faggots for eternities to come.

>RISC-V
bloated

Attached: LOW_POWER_CPU_BLENDEE.gif (400x296, 353K)

>you didn't know FPGA???
for me, it's Xilinx

how do you feel about ARM64? I have to learn ARM soon, and was gonna dive into ARM64 because I (perhaps naively) assumed it was the future. Would you recommend ARM32 instead?