skip to content
[readonly]

127 packages match "gpu"

  • total 51
  • drwxr-xr-x-../
  • drwxr-xr-x-Oct 27 03:26@ajax/mesh-prog-matmul/# 0.1.0 matmul, the mesh's exact-mode GPU GEMM and GEMV (bit-identical on every GPU), as PTX from C
  • drwxr-xr-x-Oct 27 03:26@ajax/mesh-prog-pathtrace/# 0.1.0 pathtrace, a deterministic path tracer as a mesh GPU job program (bit-identical images on every GPU and CPU), as PTX from C
  • drwxr-xr-x-Oct 27 05:24@ajax/mesh-load-share/# 0.1.0 The GPU share maths of the load sampler: node load record deltas, share, slices, gaps, and nearest-rank quantiles
  • drwxr-xr-x-Oct 27 03:26@ajax/mesh-prog-doom/# 0.1.0 doom, a DDA raycaster as a resident mesh GPU program (one lane a column, the pose and keys on the GPU, split across machines), as PTX from C
  • drwxr-xr-x-Oct 27 03:26@ajax/mesh-prog-hello/# 0.1.0 hello, the smallest mesh GPU program (task = arg * 3 + 7), as PTX from C
  • drwxr-xr-x-Oct 27 03:26@ajax/mesh-rt-sim/# 0.1.0 mesh-rt's device primitives on the CPU, and a launcher that runs a kernel's nodes as threads, for testing GPU programs without a GPU
  • drwxr-xr-x-Oct 27 03:26@ajax/mesh-rt-linalg/# 0.1.0 Exact-mode GEMM tile and GEMV at mesh node scope: sequential sums, unfused, bit-identical on every GPU
  • drwxr-xr-x-Oct 27 03:26@ajax/mesh-prog-pipehash/# 0.1.0 pipehash, the pipe demo's GPU stage: pops msgs from a CPU source, sha256d's them a lane each, pushes digests to a CPU sink, as PTX from C
  • drwxr-xr-x-Oct 27 03:26@ajax/mesh-prog-gpt2x/# 0.1.0 gpt2x, GPT-2 124M's exact-mode GPU kernels (embed, norm, GEMM/GEMV, KV, attention, logits) bitwise equal to the CPU reference, as PTX from C
  • drwxr-xr-x-Oct 27 03:26@ajax/mesh-prog-http/# 0.1.0 http, an HTTP/1.0 GET server parsed, routed and rendered by a GPU warp over the mesh program table, as PTX from C
  • drwxr-xr-x-Oct 27 02:56@ajax/cuda-driver-fake/# 0.1.0 A simulated CUDA device behind the cuda-driver table: modules, globals, streams and launches whose kernels run as C functions on CPU threads, for testing without a GPU
  • drwxr-xr-x-Oct 27 01:45@ajax/mesh-abi-view-http/# 0.1.0 programs/http's view: a GPU HTTP server's counters, load request and windows
  • drwxr-xr-x-Oct 27 03:26@ajax/mesh-rt-ptx/# 0.1.0 mesh-rt's device primitives on NVIDIA GPUs as inline PTX: special registers, sys-scope ordering, atomics, warp shuffles, the block barrier
  • drwxr-xr-x-Oct 27 01:45@ajax/mesh-abi-host-load/# 0.1.0 This machine's load (S_LOAD..): GPU share, CPU, PSI, memory, meshd, disk, thermal, cores
  • drwxr-xr-x-Oct 27 04:28@ajax/mesh-prog-sha-verify-wasm/# 0.1.0 sha-verify: re-hashes sha256d's shares on the CPU beside the GPU, reading its view through reads
  • drwxr-xr-x-Oct 27 01:45@ajax/mesh-abi-host-pload/# 0.1.0 Program load records (PL_*): GPU share, latency, throughput and CPU cost per slot
  • drwxr-xr-x-Oct 27 01:45@ajax/mesh-abi-node-load/# 0.1.0 Node load records (NL_*): how much of the GPU the nodes got, measured from inside
  • drwxr-xr-x-Oct 27 04:28@ajax/mesh-prog-pathtrace-wasm/# 0.1.0 pathtrace for meshd's wasm executor: tiles of the deterministic path tracer in the cluster float mode, bit for bit a GPU's tile
  • drwxr-xr-x-Oct 27 04:28@ajax/mesh-prog-sha256d-wasm/# 0.1.0 sha256d for meshd's wasm executor: Bitcoin-style double SHA-256 nonce grinding as a job program, the GPU image's view, ops and request language
  • drwxr-xr-x-Oct 27 04:28@ajax/mesh-prog-train-wasm/# 0.1.0 train for meshd's wasm executor: SGD of a cubic on sin(pi x) samples, lockstep with the cluster through the exchange page, bit for bit the GPUs' (cluster float mode)
  • drwxr-xr-x-Oct 27 02:57@ajax/mesh-occupancy/# 0.1.0 Co-residency of resident CUDA kernels: whether every program's nodes fit on the GPU at once, and how many more do
  • drwxr-xr-x-Oct 27 03:26@ajax/mesh-rt/# 0.1.0 The mesh device runtime: MESH_PROGRAM(init, task) makes a resident GPU kernel (mesh_main) on the node loop, or a CPU-simulated one
  • drwxr-xr-x-Oct 27 04:12@ajax/mesh-top-mock-meshd/# 0.1.0 A mock meshd control port for trying mesh-top without a GPU: synthetic snapshots, trace and log (a viewer fixture only)
  • drwxr-xr-x-Oct 27 03:26@ajax/mesh-prog-sha256d/# 0.1.0 sha256d, Bitcoin-style double SHA-256 nonce grinding as a mesh GPU job program (multi-warp nodes), as PTX from C
  • drwxr-xr-x-Oct 27 01:59@ajax/pathtrace-f32/# 0.1.0 programs/pathtrace's renderer instantiated with plain f32 operations (what the GPU computes, FTZ aside): pt_*_f32
  • drwxr-xr-x-Oct 27 07:33@ajax/meshd/# 0.1.1 meshd: the host daemon of the GPU mesh: the executors, the driver, the control and HTTP ports, the samplers, the jobs and blob threads, and a fleet member's links, leader and reconciler
  • drwxr-xr-x-Oct 27 02:56@ajax/nvjitlink/# 0.1.0 nvJitLink loaded at runtime: PTX to a cubin for the GPU's arch, off the blink, with no driver JIT fallback
  • drwxr-xr-x-Oct 27 05:24@ajax/mesh-load-prog/# 0.1.0 One program's load sample: GPU share from its node load records, task latencies from its trace, busy time, throughput, its CPU executor's cost
  • drwxr-xr-x-Oct 27 01:59@ajax/mesh-ref-matmul/# 0.1.0 programs/matmul's exact mode on the CPU: counter-hash matrices, C = A B and y = A x in the GPU's summation order, the FNV digest
  • drwxr-xr-x-Oct 27 04:12@ajax/mesh-top-ui-http/# 0.1.0 mesh-top's HTTP tab: requests the GPU served, a curl line for the bound address, the path input and the scrollable response
  • drwxr-xr-x-Oct 27 02:57@ajax/mesh-inproc/# 0.1.0 A GPU program driven in-process: compile, load, install a program-table slot in a mailbox of its own, launch, publish, stop and relaunch, through meshd's backend
  • drwxr-xr-x-Oct 27 05:23@ajax/mesh-http-relay/# 0.1.0 meshd's HTTP port: request heads relayed to the GPU http program, and the routes meshd answers itself (the views, infer, render)
  • drwxr-xr-x-Oct 27 04:12@ajax/mesh-top-ui-load/# 0.1.0 mesh-top's load panes: the Load tab (GPU share, host CPU and pressure, cores, per-program task time, share history), a machine panel's Load box and the overview's load cells
  • drwxr-xr-x-Oct 27 04:12@ajax/mesh-top-ui-mesh/# 0.1.0 mesh-top's Mesh tab: the cluster overview (a row per system, even one known only through its peers) and a panel per machine with its GPU, nodes, peers, program chips, sparkline, network, load and node grid
  • drwxr-xr-x-Oct 27 01:59@ajax/mesh-ref-block/# 0.1.0 The block collectives of a multi-warp node on the CPU: warp butterflies and scans, block reduce and scan, step for step as the GPU
  • drwxr-xr-x-Oct 27 01:45@ajax/mesh-abi-fields/# 0.1.0 Region layout checks: (offset, bytes) fields, overlap and extent
  • drwxr-xr-x-Oct 27 03:26@ajax/mesh-prog-infer/# 0.1.0 infer, the mesh program that evaluates train's cubic over a grid through a peer view, as PTX from C
  • drwxr-xr-x-Oct 27 01:59@ajax/mesh-ref-gpt2/# 0.1.0 GPT-2 124M in the cluster float mode on the CPU: the forward pass, prefill and decode, greedy generation, bitwise equal to the GPU kernels
  • drwxr-xr-x-Oct 27 01:45@ajax/mesh-abi-name/# 0.1.0 Valid program names: [a-z0-9-]{1,15}
  • drwxr-xr-x-Oct 27 03:26@ajax/mesh-rt-prim/# 0.1.0 mesh-rt's device primitives for the target being compiled (PTX on nvptx64, the CPU simulator elsewhere) and the f32 helpers on them
  • drwxr-xr-x-Oct 27 01:45@ajax/mesh-abi-pipe-hash/# 0.1.0 Pipe hashes: a record's position-mixed sum and an inbox type's FNV-1a 32
  • drwxr-xr-x-Oct 27 01:45@ajax/mesh-abi-fnv/# 0.1.0 FNV-1a 64, the hash of program images in the mesh contract
  • drwxr-xr-x-Oct 27 03:26@ajax/mesh-rt-block/# 0.1.0 Block collectives of a multi-warp mesh node in a fixed, bit-reproducible order (reduce, scan, broadcast), and typed shared memory
  • drwxr-xr-x-Oct 27 01:45@ajax/mesh-abi-fault/# 0.1.0 P_FAULT codes
  • drwxr-xr-x-Oct 27 03:26@ajax/mesh-rt-view/# 0.1.0 A bounds-checked byte window of the mailbox for mesh programs, and the program fault record
  • drwxr-xr-x-Oct 27 02:57@ajax/mesh-mps/# 0.1.0 meshd's CUDA backend: one context that blinks, or --isolation mps with a private MPS daemon and a pool of worker processes; nvJitLink as its compiler
  • drwxr-xr-x-Oct 27 01:45@ajax/mesh-abi-input-queue/# 0.1.0 A program's input queue in its view (IQ_*)
  • drwxr-xr-x-Oct 27 01:45@ajax/mesh-abi-place-ops/# 0.1.0 F_PLACE ops: pause, paused, set, done, fail
  • drwxr-xr-x-Oct 27 01:45@ajax/mesh-abi-native/# 0.1.0 The native worker contract: its ABI, environment and file descriptors
  • drwxr-xr-x-Oct 27 01:45@ajax/mesh-abi-peer-sizes/# 0.1.0 Peer frame body sizes and limits

-- more --