Wave Computing Closes Its MIPS Open Initiative with Immediate Effect
hackster.io
hackster.io
It is disappointing when a company won't publicly explain the reasons behind its decisions. It doesn't impress.
And Mips-Imagination as a commercial entity was braindead. Look for stories of them responding to sales calls after 1+ month.
Execution was below the sea level.
I can say the same of many other "Big Semi" that went nowhere in the last 5 years. They all had underperforming C-suites stuffed with overfed big name executives.
I don't think MIPS, open source or not, is competitive in a RISC-V world, and I'm not even much of a RISC-V fan (I'm more of an ARMv8 fan). Yeah, if MIPS had done this a decade ago and forestalled the rise of a RISC-V, then they may have had a chance.
The going-nowhere bureaucracy-heavy structure resonates with the impression I got from reading this - I can't help but wonder if:
- they brought in outside consultants to "strategize" how to respond to the OpenPOWER / RISC-V "threat" (postulation: RISC-V put the frozen deer in front of the headlights, while OpenPOWER was the foghorn that got them scrambling at the last moment - in any direction)
- said consultant wanted to ensure they'd get paid (translation: knew they were talking to a board that expected a yes-team), so simply coughed up "modern/relevant industry standard" advice they (probably correctly) assumed was expected of them: go open
- the entire industry went "all riiiiight, free MIPS!", and wasted no time commercially optimizing around royalty-free licencing
- MIPS' "overfed" (as you astutely put it) board suddenly saw all their holidays and etcetera flying away
- either this happened a while ago and the consultant said to hang on and it backfired, or this just happened recently
- Exceptionally loud meetings (probably involving death threats) were had, possibly angry and/or irrational "NO, WE ARE STOPPING THIS NOW" decisions were made, and...
...welp.
Given MIPS' execution being "below sea level" (again, as you put it), I get the impression the management had no resonance with the technical architecture of what it was they were managing.
So, they couldn't move the innovation forward, only industrialize the process of perpetuating what already existed.
Hence, going open without considering how to pull that off in a way that is both financially strategic and brings the best out of the architecture.
And because they didn't understand the strengths of their architecture they had no confidence/competence to execute properly in the first place, anticipate the disaster in this approach, or handle it correctly.
</naive_anecdotal_theory>
https://hackaday.com/2019/10/23/arm-allows-custom-instructio...