A modern $400 MIPS workstation for a modern (government) era (2020)
virtuallyfun.com
virtuallyfun.com
Just for yucks I ran the BYTE UNIX benchmarks 5.1.3 (same as they used) on my 4 GB VisionFive 2 (RISC-V quad core 1.5 GHz dual issue in-order CPU similar to Arm A55) which I paid $49 for on the Kickstarter.
Sadly I can't see how to correlate the results to their graph. They say they rescaled some results :-( According to the README the baseline is a SPARCstation 20-61 (60 MHz?) which is set to 10.0 for each benchmark.
System Benchmarks Index Values BASELINE RESULT INDEX
Dhrystone 2 using register variables 116700.0 5602579.7 480.1
Double-Precision Whetstone 55.0 1433.4 260.6
Execl Throughput 43.0 1832.2 426.1
File Copy 1024 bufsize 2000 maxblocks 3960.0 203687.5 514.4
File Copy 256 bufsize 500 maxblocks 1655.0 67413.7 407.3
File Copy 4096 bufsize 8000 maxblocks 5800.0 363829.1 627.3
Pipe Throughput 12440.0 391228.8 314.5
Pipe-based Context Switching 4000.0 42404.5 106.0
Process Creation 126.0 2086.5 165.6
Shell Scripts (1 concurrent) 42.4 2900.7 684.1
Shell Scripts (8 concurrent) 6.0 1253.5 2089.1
System Call Overhead 15000.0 740463.6 493.6
========
System Benchmarks Index Score 416.6
System Benchmarks Index Values BASELINE RESULT INDEX
Dhrystone 2 using register variables 116700.0 10096040.6 865.1
Double-Precision Whetstone 55.0 2389.1 434.4
Execl Throughput 43.0 927.1 215.6
File Copy 1024 bufsize 2000 maxblocks 3960.0 114719.5 289.7
File Copy 256 bufsize 500 maxblocks 1655.0 31989.5 193.3
File Copy 4096 bufsize 8000 maxblocks 5800.0 323307.3 557.4
Pipe Throughput 12440.0 162501.6 130.6
Pipe-based Context Switching 4000.0 49867.6 124.7
Process Creation 126.0 1872.7 148.6
Shell Scripts (1 concurrent) 42.4 2535.2 597.9
Shell Scripts (8 concurrent) 6.0 768.3 1280.5
System Call Overhead 15000.0 488198.2 325.5
========
System Benchmarks Index Score 327.6I think -- but do not know -- that the key thing here is that this is a Chinese made chip on a Chinese made motherboard with a Chinese chipset. And, of course, they have the source code to all the software, because it's FOSS.
It doesn't matter if the performance is great and even less the price/performance. All that matters is it works and it's usable.
So, if -- for instance -- the PRoC invaded and annexed Taiwan and all the silicon fabs in Taiwan were suddenly unusable, the rest of the world would suddenly have a big problem. No more AMD CPUs, supply problems with Arm SoCs, and RAM and flash and chipsets, so even if you could still get Intel CPUs, without readily available RAM and so on, the ROTW couldn't build cheap mass-market PCs any more.
But PRoC could.
Which makes the invasion of Taiwan a much better bet. It'd hurt China but not as much as it would hurt everyone else.
And since now the West is distracted from Ukraine by Hamas-v-Israel, Russia is getting away with it. That makes PRoC-v-RoC even more likely.
That even brings Korean War II into sight over the horizon.
:-(
Media, perhaps. No one else is. Hamas is a tiny sideshow. The Kremlin wants people to think the West is distracted.
Note that TSMC has (or is building) fabs in USA and Germany. I think an invasion of Taiwan is far less likely now than it was 2+ years ago, but if it did happen it would reduce TSMC's production capacity, not eliminate it. They'd just have to prioritise the important stuff and put iPhone production on hold.
Firstly, I think that's an artificial distinction. Secondly, I am seeing the almost total end of Russia/Ukraine discussion on the half a dozen social media sites I use, totally replaced by Palestine/Israel. (Mostly pro-Palestine, which makes me an outlier.)
So, no, I disagree. I think the distinction is incorrect and I think you're wrong anyway.
> Hamas is a tiny sideshow.
Overall I'd tend to agree. It's a regional skirmish as opposed to a former superpower making moves that could precipitate WW3.
> The Kremlin wants people to think the West is distracted.
I would rephrase that: the Kremlin wants the people in the West to be distracted. Not the same.
> Note that TSMC has (or is building) fabs in USA and Germany.
I'm well aware. They are far off and won't be ready for years.
> I think an invasion of Taiwan is far less likely now than it was 2+ years ago
I disagree; I think it's getting more urgent, for the same reasons you just gave. _Because_ there are distractions currently, _because_ there are other fabs under construction.
> but if it did happen it would reduce TSMC's production capacity, not eliminate it.
Totally wrong, according to TSMC and the Taiwanese gov't and industry observers and, well, everyone.
> They'd just have to prioritise the important stuff
Who decides? According to what?
> put iPhone production on hold.
What if that's the most profitable part? Again, who chooses?
I am sorry but I have to say I think every single point you make here is backwards and actively not just wrong but the opposite of the case.