back
user profile
rbanffy
194,611karma·63,760submissions·March 12, 2008
about
Seasoned software developer, proficient in Python, Java. Less proficient in Ruby and Lisp. A bit rusty in C and C++. Learning Erlang very slowly. Also a computer collector and restorer, lover of 8-bit computers, mainframes and interesting Unix workstations.
email: username at that google mail thing
http://about.me/rbanffy
https://linkedin.com/in/ricardobanffy
[ my public key: https://keybase.io/rbanffy; my proof: https://keybase.io/rbanffy/sigs/HtF1uAf_RNpwIkNP1-YGWP_-3doWV6S5Cc1KywXeLYo ]
recent activity (63,760 total)
comment
Why are those patches not in mainline? They appear to be generic enough not to give out any Nvidia secret sauce.
comment
Plotting that backwards shows a transistor would be the size of a continent, at least.
comment
> since 2010 slower still The increase in PFLOPS/dollar has continued accelerating, a lot from process, but also a lot by simplifying the architecture- if you had placed an H100 worth of trans…
comment
> use it like a restroom hand dryer One that runs continuously
comment
My workstations are usually tower servers, which are the same design as their 4 and 5u rack counterparts. Until the thermal management kicks in, they sound like jet planes. When the thermal management…
comment
It’s common for car companies when they enter a new market. It removes uncertainty from the second hand market. By doing that, you know upfront what the value of your used hardware will be at the time…
comment
We miss each other. But it’s ok. We’ll meet again at the great Apple Store in Heaven. In heaven, macOS is GNU-licensed. GPL3+, nothing less.
comment
His most hilarious predictions that were almost always wrong.
comment
> lotus Jazz I actually remember that.
comment
He was often wrong.
comment
Not sure whether games will run well on it.
comment
In the most realistic scenario, the movie would be a short. The debris cloud would hit the shuttle and they’d die immediately.
comment
Modern, as in “serving campaign donors rather than their voters”.
comment
The optimal case was always loading up the coprocessor memory and letting it crunch the data until it finishes. One nice thing was the low cognitive load at a time CUDA was not a great experience - it…
comment
Funny thing: modern Xeons are getting closer and closer to the core and thread counts of the Xeon Phis. A 6980P has 128 cores and 256 threads, with AVX-512 and AMX.
comment
I am surprised they have a group working on something we really don't have a definition for.
comment
A shame the x3d cache parts lack integrated GPUs. The extra cache could be handy for inference with smaller models.