varelse

94 ms·

631.
by varelse 11y ago
Methinks embrace the magical library written by engineers and researchers at Google instead of by a random goulash of machine learning grad students with no investment in the outcome other than graduating.
632.
by varelse 11y ago
FPGAs are not measuring up to GPUs yet for these tasks. There's too much floating-point math (and yes, Altera is addressing this) for forward prediction, and way too much intermediate state to save for training (GPU memory controllers
633.
by varelse 11y ago
And just 4 years after their insistence that GPUs would never have a role at Google... How 'bout that?
634.
by varelse 11y ago
The entire mobile browsing experience on Android is broken. It's slow, it's full of distracting modal ads and videos that frequently lead to #$%^ing slide shows, and to top it off, Chrome frequently crashes. And no, I don't
635.
by varelse 11y ago
And only in the US would this be immediately acted upon as such. And that's because we all have to pretend there's no cultural link between anything and anything else. For we must treat all Americans as ideal beige spherical cows
636.
by varelse 11y ago
The approach I've taken to this dilemma is to work directly with the scientists to understand the quantities they are trying to compute rather than attempt to port the code itself (yes, I actively choose refactoring over porting). Once
637.
by varelse 11y ago
I'd suggest creating a templated buffer class that manages data uploads/downloads and exposes the appropriate pointer on the appropriate device. And from an algorithmic standpoint, don't pingpong large buffers of data betwee
638.
by varelse 11y ago
s/Nanotech/Big Data/ s/Big Data/Machine Learning/ s/Machine Learning/Deep Learning/ s/Deep Learning/???/
639.
by varelse 11y ago
I think the emphasis on unified memory is a bit of a red herring here. In my experience, heterogeneous computing means optimizing data placement such that the parallel-friendly number crunching is done entirely on the GPU and I/O, pro
640.
by varelse 11y ago
AWS IMO is insanely great at cutting the price of established web services. This is more than enough to carry them for the next 5-10 years. What will be interesting is how they deal with what comes afterwards...
641.
by varelse 11y ago
No, it shouldn't. The relationship between Amazon and AWS is, um, complicated.
642.
by varelse 11y ago
Speaking as an ex-googler, I think a disturbing amount of Google is dedicated to pointless blind-allocated me-too attempts to kick sand in the face of its competitors. Of course, I also think they could work wonders on their horrible attrit
643.
by varelse 11y ago
As an example, today's cars are radically different beasts compared to what we had in 1950 (or even 1980). They're safe, faster, lighter, and full of subtle drive-by-wire algorithms and technology. And there's a huge amount
644.
by varelse 11y ago
Tomas Lindahl, uncle of Erik Lindahl, head GROMACS honcho... http://www.gromacs.org/
645.
by varelse 11y ago
I'd have given them major kudos if they'd called it Web Traffic Firewall instead.
646.
by varelse 11y ago
I'm all for optimizing the last 20%, but unless I'm mistaken, in America, ~80% of this is due to a sedentary lifestyle. And that bias arises from knowing people genetically predisposed to type 2 who have close relatives who died
647.
by varelse 11y ago
So what happens when such a person eats relatively healthily and exercises regularly? Does this do any significant good or should they just accept the inevitable?
648.
by varelse 11y ago
Type 2 is somewhat an opt-in disease, is it not? What pisses me off is how much can be done to prevent it and how little people are willing to do so, especially when their genes can tell them point blank that they really ought to be doing s
649.
by varelse 11y ago
Cool, one thing though, you're not getting my genome. That's like giving you my fingerprints and we've just barely met. My fitbit data has probably already outed me to you given what I said previously. I'm OK with tha
650.
by varelse 11y ago
Not to sound snarky, but uploading my fitbit data IMO just solved the mystery of obesity (at least for me). I burn ~3200 calories a day and I walk >10 miles a day. This is the 100th percentile of your data and that floors me. So I have
651.
by varelse 11y ago
the !/$ on ECC has been craptastic for a rather long time. In fact, on GPUs, it's more likely your GPU is bad from the get-go than it is to encounter an ECC error: http://www.robinbetz.com/papers/xsede13_betz
652.
by varelse 11y ago
You seriously need to investigate what a cluster#$%! Xeon Phi is. Their attempt to kill NVIDIA was a joke and continues to be one for anyone not getting paid to say otherwise. IMO if Intel really cared about "shareholder value(tm)&quo
653.
by varelse 11y ago
Correct, Intel is playing dirty with Skylake. Sandybridge (i7-3820), IvyBridge (i7-4820), and Haswell (i7-5930k) all had 40 PCIE lane high-end consumer CPUs. Such CPUs could be used with PLX PCIE switches to build inexpensive quad-GPU sup
654.
by varelse 11y ago
Altera has developed an OpenCL->VHDL compiler. That said, GPU code does not directly map to VHDL code in an efficient manner. Note the utter absence of GPU vs FPGA OpenCL benchmarks in the wild and take the hint... https://w
655.
by varelse 11y ago
Seems like the _Reefer Madness_ version of _The Robot Apocalypse_ to me. @#$% that Star Trek episode, Futurama IMO covered robot/human relations much better and more realistically with "I Dated A Robot": https://e
656.
by varelse 11y ago
You're taking things way too seriously. Besides, there really aren't any other practical deep learning solutions right now besides NVIDIA. And that's because AMD and Intel dropped the ball NVIDIA carried for the last 10 yea
657.
by varelse 11y ago
Sure, but most of the DNN people are just running glorified AlexNets and GoogleNets and not really looking deeply under the hood (I speak from 1st hand experience). For Caffe is open source. And cuDNN makes it fast. And if it doesn&#x27
658.
by varelse 11y ago
Reports like this are great for embarrassing NVIDIA. But the problem is that this paper is already out of date, for embarrassing NVIDIA is the best way to get them to address deficiencies like this. This is mostly already fixed: https:&#x
659.
by varelse 11y ago
One way I addressed this exact situation: Compute a hash on the state after each iteration and save it. Now run again normally and save several iterations worth of state somewhere. When the hash diverges, you have hit the race condition.
660.
by varelse 11y ago
Methinks I'm going to HN hell for presenting the disagreement here as a logical fallacy, but: https://en.wikipedia.org/wiki/Argumentum_ad_populum For I agree with drzaiuslordofthenukedfallenearthgetyourhandsoffmey

More ›