Source
- URL: https://x.com/AgileJebrim/status/2058420153348391346
- Author: Jebrim (@AgileJebrim)
- Posted: 2026-05-24 05:29:42
Thread
1/ @AgileJebrim
I’ve taken to using a simple interview question with folks. It’s surprising how much people seem to struggle with answering it.
Tell me how you would efficiently sum a large array of numbers on a GPU into a single accumulated value.
Don’t need to see code, just explain it conceptually in relation to the hardware.
Branches: 2026-05-24-stikves-they-need-to-know-how-gpu-shaders-are-organized, 2026-05-24-noinodev-i-see-it-as-a-convolution-task-gather-per-thread, 2026-05-24-saladpalad-tree-reduction, 2026-05-24-brainage19-im-bored-so-im-gonna-give-this-question-a-good, 2026-05-24-atorstling-hierarchical-map-reduce, 2026-05-24-Gombally_-i-think-this-is-a-simple-example-of-look-how-much, 2026-05-24-Lotsoffoodinme-many-people-dont-write-code-that-touches-the-gpu, 2026-05-24-ourtown2-use-a-parallel-reduction-not-atomics-from-every, 2026-05-24-tomcr2100-mapreduce-simple-enough-answer, 2026-05-24-VMises76153-i-would-embarrass-myself-by-trying-to-answer, 2026-05-24-heartoneto2012-do-floor-warpsize-arraysize-sums-keep-using-the, 2026-05-24-dadbod2100-its-for-a-frontend-javascript-position-i-see, 2026-05-24-FUZxxl-they-asked-the-same-question-to-me-at-nvidia-its, 2026-05-24-gulbaru-i-am-not-knowlegable-enough-with-gpus-but-does, 2026-05-24-feikcel-divide-into-small-arrays-conquer-each-array-on-a, 2026-05-24-MikeTreit-this-is-an-idiotic-interview-question, 2026-05-24-Jason_Ewton-m-x-n-multipass-reduction, 2026-05-24-minusYCore-does-the-array-initally-exist-in-gpu-memory-or-cpu, 2026-05-24-fatroom-no-idea-about-gpu-programming-cpu-hints-towards, 2026-05-24-2Sexy4MyGPU-lots-and-lots-of-atomics, 2026-05-24-deleted_1m-what-kind-of-gpu, 2026-05-24-kechogarcia-easy-waveactivesum-then-lds-atomic-sum-then, 2026-05-24-miru_why-efficient-reductions-are-a-very-deep-rabbit-hole, 2026-05-24-braincell__-if-you-can-fit-qwen-3-6-into-your-gpu-just-load, 2026-05-24-Apolloknius-shrink-the-array-over-and-over-until-its-down-to, 2026-05-24-matt_timmermans-divide-up-the-available-memory-between-cores-each, 2026-05-24-steverockan-answering-blind-hierarchically-split-into-chunks, 2026-05-24-CodeAndTears-large-array-of-values-is-stored-in-float-double, 2026-05-24-Lord_KX_-then-make-them-build-it-in-minecraft-using-pre, 2026-05-24-jes5199-you-shouldnt-do-that-on-the-gpu-do-it-on-the, 2026-05-24-SpikyCaterpillr-i-have-no-meaningful-gpu-coding-experience-that, 2026-05-24-uvadin-i-had-this-problem-as-an-example-in-a-uni-class, 2026-05-24-ArcticDodger-add-up-all-the-columns-separately-give-each, 2026-05-24-MichalPiotrK-if-the-array-is-in-ram-do-it-sequentially-if-its, 2026-05-24-xlrndo-what-precision-are-the-numbers-integer-or, 2026-05-25-bmcnett-when-we-were-first-evaluating-compute-shaders-at, 2026-05-25-sandovin34721-i-have-no-idea-about-gpu-hardware-and-how-it, 2026-05-25-JesseStojan-using-a-hierarchical-structured-layers-on-device, 2026-05-25-DemetriSpanos-as-a-curious-bystander-can-you-give-a-rough-idea, 2026-05-25-Karyuutensei-nvidia-has-good-resources-on-this-https-developer, 2026-05-25-Waffl3x-use-openmp-target-region-map-the-memory-use, 2026-05-25-ParkerRedford-i-would-allocate-a-variable-in-vram-then