# \#gds **URL:** https://forums.developer.nvidia.com/tag/gds/462.md [Latest](https://forums.developer.nvidia.com/latest.md) · [Categories](https://forums.developer.nvidia.com/categories.md) · [Tags](https://forums.developer.nvidia.com/tags.md) --- ## [Does GPUDirect Storage support Tesla P4 / T4 / RTX A2000](https://forums.developer.nvidia.com/t/does-gpudirect-storage-support-tesla-p4-t4-rtx-a2000/372729)
**Author:** [@1043706381](https://forums.developer.nvidia.com/u/1043706381)\ **Replies:** 1\ **Last updated:** [October 6, 2026, 12:25pm UTC](https://forums.developer.nvidia.com/t/does-gpudirect-storage-support-tesla-p4-t4-rtx-a2000/372729 "2026-10-06T12:25:50Z")
Does GPUDirect Storage support Tesla P4 / T4 / RTX A2000? --- ## [cuFile only running compat mode for RTX 6000 pro & Samsung 9100 Pro SSD on ASUS TUF Gaming B850](https://forums.developer.nvidia.com/t/cufile-only-running-compat-mode-for-rtx-6000-pro-samsung-9100-pro-ssd-on-asus-tuf-gaming-b850/362716)
**Author:** [@theaustrian](https://forums.developer.nvidia.com/u/theaustrian)\ **Replies:** 7\ **Last updated:** [August 3, 2026, 7:29am UTC](https://forums.developer.nvidia.com/t/cufile-only-running-compat-mode-for-rtx-6000-pro-samsung-9100-pro-ssd-on-asus-tuf-gaming-b850/362716 "2026-08-03T07:29:51Z")
Hi all, I’m running an RTX 6000 Blackwell (Pro Edition) on a Ryzen 7500X3D (ASUS B850 platform) with a Samsung Gen5 NVMe. I’ve successfully isolated the RTX 6000 by moving the display to the iGPU and cleaning all Xorg/… --- ## [Enabling GDS support](https://forums.developer.nvidia.com/t/enabling-gds-support/366659)
**Author:** [@alok.prasad7](https://forums.developer.nvidia.com/u/alok.prasad7)\ **Replies:** 1\ **Last updated:** [April 15, 2026, 7:13am UTC](https://forums.developer.nvidia.com/t/enabling-gds-support/366659 "2026-04-15T07:13:24Z")
I had system where GDS used to work , now running gdscheck after updating kernel it is showing NVME not supported 6.17.0-19-generic nvidia-fs loaded - 1.17.0.44 Driver Version: 590.48.01 CUDA Version: 13.1 NVIDIA H200… --- ## [DOCA NVME Patch for GDS via nvidia-fs](https://forums.developer.nvidia.com/t/doca-nvme-patch-for-gds-via-nvidia-fs/361815)
**Author:** [@alok.prasad7](https://forums.developer.nvidia.com/u/alok.prasad7)\ **Replies:** 1\ **Last updated:** [March 31, 2026, 6:45am UTC](https://forums.developer.nvidia.com/t/doca-nvme-patch-for-gds-via-nvidia-fs/361815 "2026-03-31T06:45:09Z")
I am trying to understand DOCA NVME patch required for nvidia-fs to work, based on my understanding basically we need to add 3 files in drivers/nvme/host/nvfs-dma.c drivers/nvme/host/nvfs-dma.h drivers/nvme/host/nvfs.… --- ## [How libcufile Check for O\_DIRECT ( DIO) Flag](https://forums.developer.nvidia.com/t/how-libcufile-check-for-o-direct-dio-flag/362479)
**Author:** [@alok.prasad7](https://forums.developer.nvidia.com/u/alok.prasad7)\ **Replies:** 2\ **Last updated:** [March 11, 2026, 6:20am UTC](https://forums.developer.nvidia.com/t/how-libcufile-check-for-o-direct-dio-flag/362479 "2026-03-11T06:20:42Z")
I am trying to make our device work with Libcufile , basically we want our device dma which has control of data trafer to its huge volatile memory to be used by GPu using GDS. its failing at cufio\_core:1259 cuFile DIO s… --- ## [Understanding NVME Patch for nvidia-fs support for GDS ( non p2pdma mode)](https://forums.developer.nvidia.com/t/understanding-nvme-patch-for-nvidia-fs-support-for-gds-non-p2pdma-mode/361521)
**Author:** [@alok.prasad7](https://forums.developer.nvidia.com/u/alok.prasad7)\ **Replies:** 0\ **Last updated:** [February 23, 2026, 4:08pm UTC](https://forums.developer.nvidia.com/t/understanding-nvme-patch-for-nvidia-fs-support-for-gds-non-p2pdma-mode/361521 "2026-02-23T16:08:33Z")
@kmodukuri i am trying to understand the patches required for NVIDIA-FS support for GDS, is the changes under New Files host/nvfs-dma.c host/nvfs-dma.h host/nvfs.h Patches host/pci.c(patched) Also changes under C… --- ## [Is GDS actually enabled on my system? gdscheck shows NVMe Unsupported but cuFile I/O works](https://forums.developer.nvidia.com/t/is-gds-actually-enabled-on-my-system-gdscheck-shows-nvme-unsupported-but-cufile-i-o-works/352258)
**Author:** [@pandeyshweta2401](https://forums.developer.nvidia.com/u/pandeyshweta2401)\ **Replies:** 1\ **Last updated:** [February 23, 2026, 3:58pm UTC](https://forums.developer.nvidia.com/t/is-gds-actually-enabled-on-my-system-gdscheck-shows-nvme-unsupported-but-cufile-i-o-works/352258 "2026-02-23T15:58:46Z")
Hi, I’m trying to verify whether GPUDirect Storage (GDS) is fully enabled on my system. I installed the NVIDIA Open Kernel driver, CUDA toolkit, and GDS (nvidia-fs + libcufile). gdscheck reports a supported platform, b… --- ## [NVIDIA GDS does not run PCIe P2P in xfer=0](https://forums.developer.nvidia.com/t/nvidia-gds-does-not-run-pcie-p2p-in-xfer-0/279369)
**Author:** [@snotna](https://forums.developer.nvidia.com/u/snotna)\ **Replies:** 5\ **Last updated:** [February 18, 2026, 10:49am UTC](https://forums.developer.nvidia.com/t/nvidia-gds-does-not-run-pcie-p2p-in-xfer-0/279369 "2026-02-18T10:49:55Z")
Hi, I’m trying to use Nvidia GDS to transfer data between an NVMe storage and an Nvidia GTX 3060, with no success. My setup is: PCIe RP ↔ PCIe switch ↔ Nvidia 3060 PCIe switch ↔ NVMe My setup enumerates on PCIe unde… --- ## [GDS requirement for cards](https://forums.developer.nvidia.com/t/gds-requirement-for-cards/184489)
**Author:** [@363497485](https://forums.developer.nvidia.com/u/363497485)\ **Replies:** 3\ **Last updated:** [February 11, 2026, 10:12am UTC](https://forums.developer.nvidia.com/t/gds-requirement-for-cards/184489 "2026-02-11T10:12:47Z")
Hi. I know GD RMDA is only compatible with Tesla and Quadro cards, but recently I noticed that for GDS it says Ampere Architecture are supported. Does that mean GDS works on the RTX30 series? --- ## [NVMeoF RDMA over IB intermittently fails with CQE local protection error 4 during gdsio cuFileWrite](https://forums.developer.nvidia.com/t/nvmeof-rdma-over-ib-intermittently-fails-with-cqe-local-protection-error-4-during-gdsio-cufilewrite/358989)
**Author:** [@xzhou2](https://forums.developer.nvidia.com/u/xzhou2)\ **Replies:** 1\ **Last updated:** [February 2, 2026, 5:12pm UTC](https://forums.developer.nvidia.com/t/nvmeof-rdma-over-ib-intermittently-fails-with-cqe-local-protection-error-4-during-gdsio-cufilewrite/358989 "2026-02-02T17:12:31Z")
Hi, I run into a subtle issue with gdsio write against NVMe-oF RDMA drives. It intermittently causes mlx5 driver log the following error and force nvme controller to recover, fail gdsio. \[170519.273396\] infiniband mlx… --- ## [GPU Direct Storage does not work on H20 on Ubuntu 2404 ？](https://forums.developer.nvidia.com/t/gpu-direct-storage-does-not-work-on-h20-on-ubuntu-2404/354260)
**Author:** [@panrh0904](https://forums.developer.nvidia.com/u/panrh0904)\ **Replies:** 1\ **Last updated:** [January 30, 2026, 7:40am UTC](https://forums.developer.nvidia.com/t/gpu-direct-storage-does-not-work-on-h20-on-ubuntu-2404/354260 "2026-01-30T07:40:24Z")
I have installed open-gpu-kernel-modules 580.95.05 and cuda-toolkit-13-0 on H20. When i compile nvidia-fs modules, it still generates some warnings in modpost procedure. WARNING: modpost: "nvidia\_p2p\_dma\_unmap\_pages" \[… --- ## [How to Configure and force GDS to use nvidia-fs path instead P2PDMA](https://forums.developer.nvidia.com/t/how-to-configure-and-force-gds-to-use-nvidia-fs-path-instead-p2pdma/358780)
**Author:** [@alok.prasad7](https://forums.developer.nvidia.com/u/alok.prasad7)\ **Replies:** 7\ **Last updated:** [January 29, 2026, 6:45pm UTC](https://forums.developer.nvidia.com/t/how-to-configure-and-force-gds-to-use-nvidia-fs-path-instead-p2pdma/358780 "2026-01-29T18:45:17Z")
HI , I understand while using latest kernel ( above 6.2) which has P2P DMA , we dont need to patch nvme driver/use nvidia-fs/install mofed . But what settings/config are needed to force GDS Apps to use nvidia-fs and wha… --- ## [Does Compat mode GDS uses GDRCOPY for bounce buffer](https://forums.developer.nvidia.com/t/does-compat-mode-gds-uses-gdrcopy-for-bounce-buffer/359012)
**Author:** [@alok.prasad7](https://forums.developer.nvidia.com/u/alok.prasad7)\ **Replies:** 1\ **Last updated:** [January 29, 2026, 5:38pm UTC](https://forums.developer.nvidia.com/t/does-compat-mode-gds-uses-gdrcopy-for-bounce-buffer/359012 "2026-01-29T17:38:46Z")
Hi , Does compat mode GDS uses GDRCOPY for the bounce buffers? --- ## [GDSIO Not working with T600 and samsung NVME](https://forums.developer.nvidia.com/t/gdsio-not-working-with-t600-and-samsung-nvme/358568)
**Author:** [@alok.prasad7](https://forums.developer.nvidia.com/u/alok.prasad7)\ **Replies:** 4\ **Last updated:** [January 26, 2026, 5:06pm UTC](https://forums.developer.nvidia.com/t/gdsio-not-working-with-t600-and-samsung-nvme/358568 "2026-01-26T17:06:23Z")
Hi I am trying to run GDSIO to evaluate GPUDirect Storage(GDS) Below are my config/setup Hardware NVIDIA TU117GL/T600 ( 7.5 Compute capability) NVME - Both on same NUMA Node. nvidia-smi topo -nvme NVMe0 GP… --- ## [Dose not T1000 GPU support GPU Direct Storage?](https://forums.developer.nvidia.com/t/dose-not-t1000-gpu-support-gpu-direct-storage/303217)
**Author:** [@shyeonn](https://forums.developer.nvidia.com/u/shyeonn)\ **Replies:** 1\ **Last updated:** [January 26, 2026, 5:19am UTC](https://forums.developer.nvidia.com/t/dose-not-t1000-gpu-support-gpu-direct-storage/303217 "2026-01-26T05:19:17Z")
I have a T1000 Quadro GPU and I want to use GPU direct storage API. When I write data using cufile API(Kvikio), It has error 13-08-2024 15:28:01:639 \[pid=17593 tid=17616\] ERROR 0:1534 IOCTL failed io-type 1 ret -5 ex… --- ## [What is the current proccess to contribute to GDS?](https://forums.developer.nvidia.com/t/what-is-the-current-proccess-to-contribute-to-gds/356693)
**Author:** [@j262123](https://forums.developer.nvidia.com/u/j262123)\ **Replies:** 1\ **Last updated:** [January 16, 2026, 2:56pm UTC](https://forums.developer.nvidia.com/t/what-is-the-current-proccess-to-contribute-to-gds/356693 "2026-01-16T14:56:57Z")
Hello folks, I’m a kernel developer and I have used GDS a lot. Do you know why the contribution model of GDS is just a release bump? It is not a fully OSS model where people submit PRs, someone review and then there is… --- ## [Performance Inquiry: Near-Equal Time Spent on cuFileRead()and cuFileHandleNVFS()with GDS over NFS/RDMA](https://forums.developer.nvidia.com/t/performance-inquiry-near-equal-time-spent-on-cufileread-and-cufilehandlenvfs-with-gds-over-nfs-rdma/354951)
**Author:** [@liuyuqi2001](https://forums.developer.nvidia.com/u/liuyuqi2001)\ **Replies:** 5\ **Last updated:** [December 19, 2025, 8:02am UTC](https://forums.developer.nvidia.com/t/performance-inquiry-near-equal-time-spent-on-cufileread-and-cufilehandlenvfs-with-gds-over-nfs-rdma/354951 "2025-12-19T08:02:44Z")
Hello GDS team, I’m reaching out regarding a performance observation while using DALI with GDS to load data from an NFS over RDMA. When profiling with Nsight Systems, I noticed that the time spent in the cuFileRead() c… --- ## [GDS Read/Write Error with higher IOSize - GDSIO](https://forums.developer.nvidia.com/t/gds-read-write-error-with-higher-iosize-gdsio/352851)
**Author:** [@gpmpa](https://forums.developer.nvidia.com/u/gpmpa)\ **Replies:** 0\ **Last updated:** [November 28, 2025, 3:55am UTC](https://forums.developer.nvidia.com/t/gds-read-write-error-with-higher-iosize-gdsio/352851 "2025-11-28T03:55:09Z")
Trying to setup GDS NVMe with Nvida DGX H200 server and running into errors as below while testing using gdsio with Kubernetes. Errors are appearing with high IOSize and Thread count ( here -i 2M and -w 64). gdsio -D /… --- ## [cuFileBufRegister fails above ~256 MB on RTX A6000 (48 GB HBM) with error 5016 — how to increase the maximum registration size?](https://forums.developer.nvidia.com/t/cufilebufregister-fails-above-256-mb-on-rtx-a6000-48-gb-hbm-with-error-5016-how-to-increase-the-maximum-registration-size/351408)
**Author:** [@pandeyshweta2401](https://forums.developer.nvidia.com/u/pandeyshweta2401)\ **Replies:** 0\ **Last updated:** [November 14, 2025, 10:09am UTC](https://forums.developer.nvidia.com/t/cufilebufregister-fails-above-256-mb-on-rtx-a6000-48-gb-hbm-with-error-5016-how-to-increase-the-maximum-registration-size/351408 "2025-11-14T10:09:29Z")
Hello, I am using GPUDirect Storage (GDS) on a machine with an NVIDIA RTX A6000 (48 GB HBM) and Ubuntu 24.04 with kernel 6.8.0-85-generic. I am trying to pre-register a large GPU buffer with cuFileBufRegister() for use … --- ## [How to Compile NVMe Modules with DOCA Installation](https://forums.developer.nvidia.com/t/how-to-compile-nvme-modules-with-doca-installation/332754)
**Author:** [@yk1234](https://forums.developer.nvidia.com/u/yk1234)\ **Replies:** 3\ **Last updated:** [November 14, 2025, 4:47am UTC](https://forums.developer.nvidia.com/t/how-to-compile-nvme-modules-with-doca-installation/332754 "2025-11-14T04:47:21Z")
Hi, I’m running into an issue related to NVMe-over-RDMA after installing DOCA. Previously, I installed MLNX\_OFED manually using: sudo ./mlnxofedinstall --with-nvmf --with-nfsrdma --enable-gds --add-kernel-support --dkm… --- ## [\[Kernel BUG in GDS on DGX OS6\] \`nvidia-fs:nvfs\_pin\_gpu\_pages:1383 Incompatible page table version 0x00020000\`](https://forums.developer.nvidia.com/t/kernel-bug-in-gds-on-dgx-os6-nvidia-fs1383-incompatible-page-table-version-0x00020000/350209)
**Author:** [@user157267](https://forums.developer.nvidia.com/u/user157267)\ **Replies:** 8\ **Last updated:** [November 10, 2025, 1:25pm UTC](https://forums.developer.nvidia.com/t/kernel-bug-in-gds-on-dgx-os6-nvidia-fs1383-incompatible-page-table-version-0x00020000/350209 "2025-11-10T13:25:23Z")
Hi GDS team, This is essentially a copy of an issue GPUDirect Storage read fails with kernel error · Issue #66 · NVIDIA/gds-nvidia-fs · GitHub in the nvidia-fs driver repo, but I wanted to reach out here as well, since… --- ## [CUDA 13.0 won't work with open-source drivers (NVIDIA H100 NVL)](https://forums.developer.nvidia.com/t/cuda-13-0-wont-work-with-open-source-drivers-nvidia-h100-nvl/344079)
**Author:** [@mpeim](https://forums.developer.nvidia.com/u/mpeim)\ **Replies:** 2\ **Last updated:** [November 4, 2025, 8:09pm UTC](https://forums.developer.nvidia.com/t/cuda-13-0-wont-work-with-open-source-drivers-nvidia-h100-nvl/344079 "2025-11-04T20:09:10Z")
Hi, I am currently trying to setup a lab with 4 H100 NVL to have GPUDirect Storage to local and remote NVMe drive. I have follow CUDA and NVIDIA drivers installation guide: $ nvidia-smi Thu Sep 4 22:33:04 2025 +-----… --- ## [cuFileBatchIO fails with mixed IOs](https://forums.developer.nvidia.com/t/cufilebatchio-fails-with-mixed-ios/347327)
**Author:** [@xzhou2](https://forums.developer.nvidia.com/u/xzhou2)\ **Replies:** 0\ **Last updated:** [October 10, 2025, 3:33pm UTC](https://forums.developer.nvidia.com/t/cufilebatchio-fails-with-mixed-ios/347327 "2025-10-10T15:33:47Z")
During testing with cuFileBatchIO, I encountered a cuFile I/O error. The test involves two batchIO requests, each writing six 256KiB data blocks from six separate GPU buffers to NVMe-oF backend drives. The first batch… --- ## [Io\_limit too small](https://forums.developer.nvidia.com/t/io-limit-too-small/250161)
**Author:** [@ExtremeViscent](https://forums.developer.nvidia.com/u/ExtremeViscent)\ **Replies:** 5\ **Last updated:** [October 9, 2025, 7:23pm UTC](https://forums.developer.nvidia.com/t/io-limit-too-small/250161 "2025-10-09T19:23:18Z")
Hi, During my attempts using GDS batch API, I found that the supported size for each operation is too small. When I tried to write a file larger than 100KB, I got the error: 30-03-2023 02:34:07:294 \[pid=29229 tid=2922… --- ## [\[GDS\] Per-SSD Statistics for Profiling and Monitoring](https://forums.developer.nvidia.com/t/gds-per-ssd-statistics-for-profiling-and-monitoring/346406)
**Author:** [@user157267](https://forums.developer.nvidia.com/u/user157267)\ **Replies:** 0\ **Last updated:** [September 29, 2025, 1:16pm UTC](https://forums.developer.nvidia.com/t/gds-per-ssd-statistics-for-profiling-and-monitoring/346406 "2025-09-29T13:16:32Z")
Hi GDS team, I have a quick question regarding GDS profiling and performance monitoring. I’m working with a DGX-A100 setup that includes 8 GPUs and multiple SSDs. Currently, I’m able to monitor global GDS activity acros… --- ## [GDS read/write error:use gdsio -I 1 to write](https://forums.developer.nvidia.com/t/gds-read-write-error-use-gdsio-i-1-to-write/345198)
**Author:** [@271732480](https://forums.developer.nvidia.com/u/271732480)\ **Replies:** 4\ **Last updated:** [September 22, 2025, 10:27am UTC](https://forums.developer.nvidia.com/t/gds-read-write-error-use-gdsio-i-1-to-write/345198 "2025-09-22T10:27:45Z")
Hello, I installed gds on ubuntu20.04 and it seems to be correct.Here is my gdscheck -p result: GDS release version: 1.14.1.1 nvidia\_fs version: 2.25 libcufile version: 2.12 Platform: x86\_64 ============ ENVIRONM… --- ## [Nvidia GDS issue on HGX A100 - Dell PowerEdge XE9680](https://forums.developer.nvidia.com/t/nvidia-gds-issue-on-hgx-a100-dell-poweredge-xe9680/293621)
**Author:** [@luciano.vieira](https://forums.developer.nvidia.com/u/luciano.vieira)\ **Replies:** 12\ **Last updated:** [September 17, 2025, 1:39am UTC](https://forums.developer.nvidia.com/t/nvidia-gds-issue-on-hgx-a100-dell-poweredge-xe9680/293621 "2025-09-17T01:39:19Z")
I am experiencing a problem when attempting to write to an NVME (EXT4) SSD local disk. Below are the details of the issue: Read Test: Command: /usr/local/cuda-12.4/gds/tools/gdsio -f /home/m324528/mnt\_nvme3/reg1G -d 4… --- ## [Does RTX PRO 4500 Blackwell support GPU Direct Storage?](https://forums.developer.nvidia.com/t/does-rtx-pro-4500-blackwell-support-gpu-direct-storage/344663)
**Author:** [@pinjayyu](https://forums.developer.nvidia.com/u/pinjayyu)\ **Replies:** 0\ **Last updated:** [September 11, 2025, 3:50am UTC](https://forums.developer.nvidia.com/t/does-rtx-pro-4500-blackwell-support-gpu-direct-storage/344663 "2025-09-11T03:50:22Z")
Before making a purchase, I’d like to confirm whether the RTX PRO 4500 Blackwell supports GPU Direct Storage. According to the official documentation, RTX professional desktop products support GPU Direct Storage, but I w… --- ## [Why does GDS write performance degrade when increasing the number of threads?](https://forums.developer.nvidia.com/t/why-does-gds-write-performance-degrade-when-increasing-the-number-of-threads/344034)
**Author:** [@user161369](https://forums.developer.nvidia.com/u/user161369)\ **Replies:** 0\ **Last updated:** [September 4, 2025, 1:03pm UTC](https://forums.developer.nvidia.com/t/why-does-gds-write-performance-degrade-when-increasing-the-number-of-threads/344034 "2025-09-04T13:03:12Z")
Hi, I’m running experiments using gdsio to measure GPU Direct Storage (GDS) performance. I observed that the write throughput decreases when I increase the number of threads. Specifically: Throughput increases as th… --- ## [Gdscheck says gds supported but nvidiafs shows error](https://forums.developer.nvidia.com/t/gdscheck-says-gds-supported-but-nvidiafs-shows-error/343712)
**Author:** [@nkvish105](https://forums.developer.nvidia.com/u/nkvish105)\ **Replies:** 2\ **Last updated:** [September 2, 2025, 3:26pm UTC](https://forums.developer.nvidia.com/t/gdscheck-says-gds-supported-but-nvidiafs-shows-error/343712 "2025-09-02T15:26:31Z")
Using an Ubuntu 24.04 , linux kernel 6.8.0-48-generic and nvme SSD with NVIDIA Corporation GA107M \[GeForce RTX 3050 Ti Mobile\] GPU . using Driver Version: 570.172.08 and CUDA Version: 12.8. This is my gdscheck output G… [Next page](https://forums.developer.nvidia.com/tag/gds/462.md?match_all_tags=true&page=1&tags%5B%5D=gds)