HackerReposA FIELD GUIDE TO OPEN SOURCE
ProSponsor the shelf♧ 0
← Back to top all-time
xaskasdf
ntransformer
C++2026–2026
xaskasdf / ntransformer

ntransformer

High-efficiency LLM inference engine in C++/CUDA. Run Llama 70B on RTX 3090.

AI & Machine LearningSystems & Infrastructure
Explore on GitHubRead Hacker News discussions
2Distinct people
2Mentions
0Explicit recommenders
469GitHub stars
REPOSITORY AT A GLANCE

What to know before exploring

What it is

High-efficiency LLM inference engine in C++/CUDA. Run Llama 70B on RTX 3090.

GitHub

Where it is used

No documented product use has been verified yet.

What people are saying

No recommendation phrase was found in the indexed comments.

No classified warning is attached to this page. Read the discussions before treating silence as approval.

Also in these discussions

No other repository is repeatedly linked in the same discussions.

This week on Hacker News

Nov 7–Nov 14 UTC · Direct repository links and separately labeled related project stories. Indexed through 11/14/2026, 3:56:07 PM UTC.

No direct repository links were found in the indexed HN discussions over the past seven days. Older discussions are available below.

Mentions over time

HN mentions by 6-hour interval2 dated mentions

Observed mentions: 2026-02-21 20:57:30 UTC to 2026-02-28 08:10:36 UTC.

012026-02-21 18:00:00 UTC to 2026-02-22 00:00:00 UTC (exclusive): 1 mentions; 1 submissions, 0 commentsFeb 21, 18:002026-02-22 00:00:00 UTC to 2026-02-22 06:00:00 UTC (exclusive): 0 mentions; 0 submissions, 0 comments2026-02-22 06:00:00 UTC to 2026-02-22 12:00:00 UTC (exclusive): 0 mentions; 0 submissions, 0 comments2026-02-22 12:00:00 UTC to 2026-02-22 18:00:00 UTC (exclusive): 0 mentions; 0 submissions, 0 comments2026-02-22 18:00:00 UTC to 2026-02-23 00:00:00 UTC (exclusive): 0 mentions; 0 submissions, 0 comments2026-02-23 00:00:00 UTC to 2026-02-23 06:00:00 UTC (exclusive): 0 mentions; 0 submissions, 0 comments2026-02-23 06:00:00 UTC to 2026-02-23 12:00:00 UTC (exclusive): 0 mentions; 0 submissions, 0 comments2026-02-23 12:00:00 UTC to 2026-02-23 18:00:00 UTC (exclusive): 0 mentions; 0 submissions, 0 comments2026-02-23 18:00:00 UTC to 2026-02-24 00:00:00 UTC (exclusive): 0 mentions; 0 submissions, 0 comments2026-02-24 00:00:00 UTC to 2026-02-24 06:00:00 UTC (exclusive): 0 mentions; 0 submissions, 0 comments2026-02-24 06:00:00 UTC to 2026-02-24 12:00:00 UTC (exclusive): 0 mentions; 0 submissions, 0 comments2026-02-24 12:00:00 UTC to 2026-02-24 18:00:00 UTC (exclusive): 0 mentions; 0 submissions, 0 comments2026-02-24 18:00:00 UTC to 2026-02-25 00:00:00 UTC (exclusive): 0 mentions; 0 submissions, 0 comments2026-02-25 00:00:00 UTC to 2026-02-25 06:00:00 UTC (exclusive): 0 mentions; 0 submissions, 0 commentsFeb 25, 00:002026-02-25 06:00:00 UTC to 2026-02-25 12:00:00 UTC (exclusive): 0 mentions; 0 submissions, 0 comments2026-02-25 12:00:00 UTC to 2026-02-25 18:00:00 UTC (exclusive): 0 mentions; 0 submissions, 0 comments2026-02-25 18:00:00 UTC to 2026-02-26 00:00:00 UTC (exclusive): 0 mentions; 0 submissions, 0 comments2026-02-26 00:00:00 UTC to 2026-02-26 06:00:00 UTC (exclusive): 0 mentions; 0 submissions, 0 comments2026-02-26 06:00:00 UTC to 2026-02-26 12:00:00 UTC (exclusive): 0 mentions; 0 submissions, 0 comments2026-02-26 12:00:00 UTC to 2026-02-26 18:00:00 UTC (exclusive): 0 mentions; 0 submissions, 0 comments2026-02-26 18:00:00 UTC to 2026-02-27 00:00:00 UTC (exclusive): 0 mentions; 0 submissions, 0 comments2026-02-27 00:00:00 UTC to 2026-02-27 06:00:00 UTC (exclusive): 0 mentions; 0 submissions, 0 comments2026-02-27 06:00:00 UTC to 2026-02-27 12:00:00 UTC (exclusive): 0 mentions; 0 submissions, 0 comments2026-02-27 12:00:00 UTC to 2026-02-27 18:00:00 UTC (exclusive): 0 mentions; 0 submissions, 0 comments2026-02-27 18:00:00 UTC to 2026-02-28 00:00:00 UTC (exclusive): 0 mentions; 0 submissions, 0 comments2026-02-28 00:00:00 UTC to 2026-02-28 06:00:00 UTC (exclusive): 0 mentions; 0 submissions, 0 comments2026-02-28 06:00:00 UTC to 2026-02-28 12:00:00 UTC (exclusive): 1 mentions; 0 submissions, 1 commentsFeb 28, 06:00

Counts come from the repository’s dated source evidence. Index cutoff: 2026-11-02 08:16:59 UTC.

View mention counts and date ranges
UTC ranges include their start and exclude their end. Select a period to filter the source evidence.
PeriodStart (UTC)End (exclusive, UTC)MentionsSubmissionsComments
Feb 21, 18:002026-02-21 18:00:00 UTC2026-02-22 00:00:00 UTC110
Feb 22, 00:002026-02-22 00:00:00 UTC2026-02-22 06:00:00 UTC000
Feb 22, 06:002026-02-22 06:00:00 UTC2026-02-22 12:00:00 UTC000
Feb 22, 12:002026-02-22 12:00:00 UTC2026-02-22 18:00:00 UTC000
Feb 22, 18:002026-02-22 18:00:00 UTC2026-02-23 00:00:00 UTC000
Feb 23, 00:002026-02-23 00:00:00 UTC2026-02-23 06:00:00 UTC000
Feb 23, 06:002026-02-23 06:00:00 UTC2026-02-23 12:00:00 UTC000
Feb 23, 12:002026-02-23 12:00:00 UTC2026-02-23 18:00:00 UTC000
Feb 23, 18:002026-02-23 18:00:00 UTC2026-02-24 00:00:00 UTC000
Feb 24, 00:002026-02-24 00:00:00 UTC2026-02-24 06:00:00 UTC000
Feb 24, 06:002026-02-24 06:00:00 UTC2026-02-24 12:00:00 UTC000
Feb 24, 12:002026-02-24 12:00:00 UTC2026-02-24 18:00:00 UTC000
Feb 24, 18:002026-02-24 18:00:00 UTC2026-02-25 00:00:00 UTC000
Feb 25, 00:002026-02-25 00:00:00 UTC2026-02-25 06:00:00 UTC000
Feb 25, 06:002026-02-25 06:00:00 UTC2026-02-25 12:00:00 UTC000
Feb 25, 12:002026-02-25 12:00:00 UTC2026-02-25 18:00:00 UTC000
Feb 25, 18:002026-02-25 18:00:00 UTC2026-02-26 00:00:00 UTC000
Feb 26, 00:002026-02-26 00:00:00 UTC2026-02-26 06:00:00 UTC000
Feb 26, 06:002026-02-26 06:00:00 UTC2026-02-26 12:00:00 UTC000
Feb 26, 12:002026-02-26 12:00:00 UTC2026-02-26 18:00:00 UTC000
Feb 26, 18:002026-02-26 18:00:00 UTC2026-02-27 00:00:00 UTC000
Feb 27, 00:002026-02-27 00:00:00 UTC2026-02-27 06:00:00 UTC000
Feb 27, 06:002026-02-27 06:00:00 UTC2026-02-27 12:00:00 UTC000
Feb 27, 12:002026-02-27 12:00:00 UTC2026-02-27 18:00:00 UTC000
Feb 27, 18:002026-02-27 18:00:00 UTC2026-02-28 00:00:00 UTC000
Feb 28, 00:002026-02-28 00:00:00 UTC2026-02-28 06:00:00 UTC000
Feb 28, 06:002026-02-28 06:00:00 UTC2026-02-28 12:00:00 UTC101

What Hacker News says

2 sources
xaskasdf·Feb 21, 2026Submission

Show HN: Llama 3.1 70B on a single RTX 3090 via NVMe-to-GPU bypassing the CPU

Hi everyone, I'm kinda involved in some retrogaming and with some experiments I ran into the following question: "It would be possible to run transformer models bypassing the cpu/ram, connecting the gpu to the nvme?" This is the result of that question itself and some weekend vibecoding (it has the linked library repository in the readme as well), it seems to work, even on consumer gpus, it should work better on professional ones tho
395 points · Open discussion

MAKE YOURSELF AT HOME

A few shortcuts

/
Search the collection
?
Open this guide
Esc
Close dialogs and navigation

Your saved repositories and interests stay in this browser.