rrss

11 ms·

91.
by rrss 4y ago
do you mean configuring the other PMUs to interrupt the core every function call / return? If yes, then yes that is much much higher overhead than processor trace.
92.
by rrss 4y ago
Taiwan
93.
by rrss 4y ago
does bors limit the commit frequency / PR merge frequency to one merge every test runtime (5 minutes in OP), or does it do something more complex and speculate / run some stuff in parallel? OP’s numbers are 120 commits per hour an
94.
by rrss 4y ago
Being able to just buy something that works when you get it is a big plus IMO. I’ve tried a few times in the last year and finding parts (that are actually available) for a sub- $500 small-form-factor system seems to be beyond my capabiliti
95.
by rrss 4y ago
> are any companies other than Nvidia embarking up AI/ML chipmaking in a closed fashion? https://www.cadence.com/en_US/home/solutions/machine-learnin... https://www.synopsys.com/implem
96.
by rrss 4y ago
which openroad tools use "AI" rather than normal optimization like traditional EDA tools have used for decades?
97.
by rrss 4y ago
> Btw, as far as I can tell, TPUs are really just a variant of GPUs with an emphasis on lower precision? GPUs and TPUs have substantially different architectures, it’s not accurate to consider a TPU an a variant of a GPU. See https:&#x2
98.
by rrss 5y ago
The only info I can find on people who were charged for those retail thefts who have already been sentenced and released are these cases where the DA reduced the felony charges for a misdemeanor plea deal. https://abc7news.com&#x
99.
by rrss 5y ago
something I’ve wondered for a while: Suppose someone invents something and uses it for their products, but never publishes anything about the invention, and then someone else (years later) independently develops the same thing and gets a pa
100.
by rrss 5y ago
> for example the P100 has hardware scheduling, where as the 1080 does not (in the same way at least). what does hardware scheduling mean in this context?
101.
by rrss 5y ago
because many times people want to be able to run new drivers (e.g. with new hardware support) on an LTS kernel. lots of vendors support this, and most do it by maintaining a version of their driver that can work with a wide range of kernel
102.
by rrss 5y ago
removed
103.
by rrss 5y ago
I do not understand how self directed learning like this is expected to work at all for children. If you replace a high school with letting the students do whatever interests them, a small fraction might study something, and the majority wi
104.
by rrss 5y ago
regarding prop 22's 7/8ths, i commented on this at the time: https://news.ycombinator.com/item?id=24867819 IIRC, the comment to which I replied reflected a pretty common understanding of the issue and was in some
105.
by rrss 5y ago
> If you have to ask > Only if you’re an amateur
106.
by rrss 5y ago
The first two risks are only unchanged for a demonstration that was not announced to japan, and IMO a demonstration of a nuclear bomb without telling Japan to observe it would not have been effective. After the bombing of Hiroshima, Suzuki
107.
by rrss 5y ago
> The only thing this patent covers is the physical circuit implementation The claims in that patent are not limited to a physical circuit implementation. The description includes a circuit diagram as one possible embodiment of the claim
108.
by rrss 5y ago
yes, this is linked from the article
109.
by rrss 5y ago
yes, you are right. meaningful voluntary collective change is unlikely. however, i believe that individuals should respect the commons, and don’t believe in the tragedy of the commons as an excuse for individual actions that contribute to t
110.
by rrss 5y ago
the world would be better if more companies were able or daring enough to fight bullshit patents directly (with prior art, obviousness, etc defenses). but instead every company chooses to defend themselves by deluging the patent office with
111.
by rrss 5y ago
I disagree. Not all the claims include the “single instruction cycle,” and IMO several of the claims are general enough that one could make a case that a sequence of instructions directing the operation of a general purpose data path would
112.
by rrss 5y ago
It seems like this is basically unavoidable on existing hardware, though, right? if we imagine there existed some visualization that could more accurately represent the complexity of a core, I don’t know how it would be possible to get the
113.
by rrss 5y ago
lol thanks for the treatise on the consequences of a few more days delay, but all I’ve said is that I find it amusing that the launch has ended up on Christmas day. They could have conceivably tried to avoid the holidays, but obviously they
114.
by rrss 5y ago
my personal preference is irrelevant. my comment was regarding whether it is tractable to plan around Christmas, or unreasonable because there are too many holidays of comparable significance.
115.
by rrss 5y ago
I did not get your real or feigned ignorance that Christmas is a much more significant holiday to the vast majority of the people involved than the Feast of the Ass. I also hope all goes well. Cheers.
116.
by rrss 5y ago
Cool. Webb will be monitored and operated by NASA in Maryland. https://en.m.wikipedia.org/wiki/James_Webb_Space_Telescope#T... https://en.m.wikipedia.org/wiki/Space_Telescope_Science_Inst...
117.
by rrss 5y ago
They did though. It was delayed, they chose to delay it from October to December 18 to December 22 to Christmas instead of like the first week of January or whatever. It’s not like this thing is launching to Mars and they’d have to wait a c
118.
by rrss 5y ago
How about planning around the single most celebrated holiday in the US, religious or not: https://en.m.wikipedia.org/wiki/Public_holidays_in_the_Unite...
119.
by rrss 5y ago
> It wasn’t originally planned for tomorrow Right, it was originally planned for 2007. It is sorta funny that after 14 years of delays they picked the week of Christmas.
120.
by rrss 5y ago
It looks like it was even before Falcon 1’s first success.

More ›