Tachyum
 

Anyone heard of these guys? They just came out of stealth mode in 2017 (I think)



[url]https://www.tachyum.com/[/url]

[QUOTE=aurashift;521684]Anyone heard of these guys? They just came out of stealth mode in 2017 (I think)
[url]https://www.tachyum.com/[/url][/QUOTE]Never heard of them. Could be very interesting, could be doomed because their stuff isn't x86-compatible and/or insufficiently astoundingly faster that what Intel /AMD / (ARM licensees) can provide.

Their tech spec sheets on the web are somewhat uninformative which makes me rather suspicious.

As far as I can tell, there is no FPU, unless that is part of the TPU.

Sounds way too good to be true, beating both Intel and ARM in a brand new chip?

Sounds just as far fetched as unknown people claiming to have proven the Riemann hypothesis.

[url]https://www.extremetech.com/computing/294927-tachyum-raises-25m-for-universal-processor-faster-than-xeon-smaller-than-arm[/url]

[url]https://www.anandtech.com/show/13255/hot-chips-2018-tachyum-prodigy-cpu-live-blog[/url]

[QUOTE=ATH;521699]Sounds way too good to be true, beating both Intel and ARM in a brand new chip?

Sounds just as far fetched as unknown people claiming to have proven the Riemann hypothesis.

[URL]https://www.extremetech.com/computing/294927-tachyum-raises-25m-for-universal-processor-faster-than-xeon-smaller-than-arm[/URL]

[URL]https://www.anandtech.com/show/13255/hot-chips-2018-tachyum-prodigy-cpu-live-blog[/URL][/QUOTE]


I was at that Hot Chips presentation last year and it definitely sounded too good to be true. But I'm definitely in favor of at least [I]somebody[/I] trying to dethrone x86.


OTOH, the Riemann Hypothesis has plenty of attackers at one moment in time...

[QUOTE=Mysticial;521707]I was at that Hot Chips presentation last year and it definitely sounded too good to be true.[/QUOTE]
Many too good to be true chips have been announced in the last years. Alas nothing has materialized yet. I think it's sad because we need some fresh air in fast and efficient micro architecture.



[QUOTE] But I'm definitely in favor of at least [I]somebody[/I] trying to dethrone x86.[/QUOTE]
ARM is definitely trying :smile:

"Development tools: FPGA Emulator in December 2019".

I would wait at least 3-4 years before thinking about them.

There is also the [url=https://millcomputing.com/]Mill Computing[/url] CPU.

Their claim is "Faster, Cooler, Safer". The "Cooler" means power efficiency, although the other common meaning of great/awesome is perhaps implied also I guess.

[QUOTE=ATH;521699]
Sounds just as far fetched as unknown people claiming to have proven the Riemann hypothesis.
[/QUOTE]
Any reading on how GIMPS work is helpful on that matter?

[QUOTE=retina;521721]There is also the [URL="https://millcomputing.com/"]Mill Computing[/URL] CPU.

Their claim is "Faster, Cooler, Safer". The "Cooler" means power efficiency, although the other common meaning of great/awesome is perhaps implied also I guess.[/QUOTE]
Thanks, I couldn't remember that name! What a joke they are. Incredible claims and nothing to prove after more than 10 years. I don't think they ever demoed their supposed implementation on an FPGA....


Another company with a new micro architecture, named Soft Machines, was acquired by Intel, so I guess they had something interesting to propose


[url]https://www.anandtech.com/show/10025/examining-soft-machines-architecture-visc-ipc[/url]

[QUOTE=ldesnogu;521723]Another company with a new micro architecture, named Soft Machines, was acquired by Intel, so I guess they had something interesting to propose
[/QUOTE]Another is [URL="https://en.wikipedia.org/wiki/Adapteva"]Adapteva[/URL] with their Parallela architecture. At least they shipped kit after a successful Kickstarter campain. A 4-board cluster lives on my desk in the UK; it's not been powered up for ages, largely because the USB port doesn't work after at least 75% of the reboots.