Posts
Chat
Members
Info
An IR which returns a wrong answer.
I construct a test which uses a vector containing struct. But the length of result is 0. When I change the input data to i32 vector, the length of result is correct. Blow is the code. //Multi Column Test let code = "|v: vec[{i32,i32,i32}] | …
0
0
An active weld community
I am very interested in this project and have already forked this project. Since the previous community is no longer active, I will work on the new weld community in the future. I have a preliminary understanding of this project, but I may not understand it correctly. I welcome…
0
0
·100d# General
So many questions.
When I checked the benchmark tests of the project, I found that none of these tests seemed to be available. I don't know if you can use it. After modifying some parts, I found that under the Crime_Index test case, the performance did not improve but decreased. It seems that you…
1
5
Is the project still active?
Hello everyone, I am very impressed about the work done in Weld, and I am even more thrilled about the usage of Rust / LLVM to optimize big data workloads. Still, I am willing to contribute or extend the project, but it seems the project is in a hiatus since April 2020. Have…
0
5
·100d# General
Grizzly performance degradation problem
I don’t know if you have tested this example: https://github.com/weld-project/weld/tree/master/examples/python/grizzly/get_population_stats_grizzly.py and get_population_stats_simplified_grizzly.py. In these two examples, performance has not improved. Is the method I tested…
1
11
Activity?
Hello all, I just stumbled upon weld and am very excited about its potential. There doesn't seem to have been any activity for a while (since April). I imagine Covid-19 may have had something to do with it but is the project still active?
1
1
Documentation for creating a new hardware backend?
As I see it the possible backends are LLVM (x86 supported) or openCL (for gpus). I've been tinkering with a (now deprecated) xeonphi coprocessor, with the knc (knight's corner) instruction set. LLVM doesn't support this family of x86 because of it's strange 512-bit pipeline…
1
0
Rust, weld, R and parallelism
I've watched the Weld-talk online, and I am partially hooked. Recently, I've spent a lot of time and effort into learning rust, and I am thus far impressed with it as a programming language, but I wish to understand the landscape of scientific computing a bit more. What are the…
0
0
Runtime struct packing to byte array and typecasting slices
Is it possible to generalise the input data to accept a byte array of a manually packed struct and typecast individual slices back to original type from weld?
0
3