Viewing profile — jhj
jhj
HN member- Joined
- Sat, Jun 14, 2014, 3:33 AM UTC
- HN karma
- 576
- Public activity
- 135 items
- HN profile
- View on Hacker News ↗
About jhj
https://github.com/wickedfoo
Recent public activity
-
comment
Comment #47429528
This might include people working in lumber camps in places like Siberia, "mercenaries" in Ukraine, people in NK-managed restaurants in China, Laos etc, or similar efforts that hav…
-
comment
Comment #46611521
These flops are not the same. The 2013 phone flops are fp32, the A13 flops look to be fp32 as well (not entirely sure), while the Cray numbers (like the rest of the HPC industry) a…
-
comment
Comment #43799466
Unlike quantization, dimensionality reduction/low rank approximation, distillation etc, lossless compression is an always-correct addition to any ML system as you are computing the…
-
comment
Comment #43799439
Not really, it's just adding some data transposition (coalescing individual bytes from the data words together) and an option to use a LZ/dictionary-type compressor to compress red…
-
comment
Comment #43798339
This is just a consequence of the fact that bfloat16 has a very high dynamic range which is not all used. People like hyperparameters that look like 0.01 not 10^10, even though the…
-
comment
Comment #43031292
re #3, if your RSU windfall is substantially large, you might be eligible for the 100%/110% safe harbor that won't penalize you for tax underpayments (assuming you are a US taxpaye…
-
comment
Comment #42992857
I have some of (possibly the?) cheapest residential electric power in the US, at 5.58 cents per kWh all-in cost here in Wyoming, 90%+ hydropower. Absolute lowest cold here each yea…
-
comment
Comment #42805874
Brute-force indices are usually arithmetic bound (e.g., GEMM). Cell-probe based indices are usually memory bandwidth bound (IVF, LSH bucketing, etc). Graph-based indices are usuall…
-
comment
Comment #42055968
If you have a limited number of long range ICBMs then you will likely prefer more directly military targets rather than a manufacturing facility which would likely only start to ma…
-
comment
Comment #41895521
As someone who has worked in this space (approximate compute) on both GPUs and in silicon in my research, the power consumption claims are completely bogus, as are the accuracy cla…
-
comment
Comment #41815552
> The first thing to consider is the register pressure. Increasing the number of registers per thread to optimize for ILP can lead to register spilling when the register file is ex…
-
comment
Comment #41809713
Aiming for higher occupancy is not always a desired solution, what frequently matters more is avoiding global memory latencies by retaining more data in registers and/or shared mem…
-
comment
Comment #41702195
Remote start is accidental carbon monoxide poisoning waiting to happen if your garage is directly connected to your residence. I live in an area with brutal winters in Wyoming and …
-
comment
Comment #41241332
The original sin here is that original 1980s designs carry over: the processor retains FP unit state, rather than each instruction indicating what subnormal flush mode (or rounding…
-
comment
Comment #40814572
A less risky use is to use the model to choose compilation flags and pass orderings many (but not all) of which (in theory) should always be correct but that's more of a problem of…
-
comment
Comment #40674821
While the median is much much lower, there are a couple of thousand individual contributor SWEs (non-managers) between Google, Meta and a few other big-ish tech companies who make …
-
comment
Comment #40260568
Construction costs here (Teton County, WY) are significantly higher than CA or most places in the US due to labor constraints (we have the highest average per capita income in the …
-
comment
Comment #40260413
It’s probably more California regulations than the wildfire risk per se? The direct backyard of my house in Wyoming is Bridger-Teton National Forest, wooded mountainous wilderness …
-
comment
Comment #40008589
Not just MPI over a network. We can compress floats, send them over NVLink or PCIe to another GPU in the same host, and decompress and it can be faster than sending data raw betwee…
-
comment
Comment #40006752
People in the HPC/classical supercomputing space have done this sort of thing for a while. There's a fair amount of literature on lossless floating point compression, such as Marti…
-
comment
Comment #39866541
I live near Yellowstone in Wyoming. The park is a lot more massive than Zion, usually involving multi-hour drives to get around, and there are multiple roads in the park, all of wh…
-
comment
Comment #39720860
Finite field log/antilog lookup tables are used for efficient-ish multiplication, similar to addition/subtraction tables used for logarithmic number systems.
-
comment
Comment #39720475
> By scaling the numbers first by taking the log, multiplication becomes addition and addition becomes x + log1p(exp(y - x)). Addition/subtraction in a logarithmic number system is…
-
comment
Comment #39146625
There is no good geometry to be exploited, and the query vectors might be (and are usually) distributed quite differently than the indexed vectors. For Euclidean (L2) distance inde…
-
comment
Comment #39143639
Speaking as an author of one of the primary libraries for doing this stuff (faiss), it is not because it is still an open ended research problem on how approximate high-dimensional…