#1569 AMD/HIP RX 9070 XT (gfx1201), Ubuntu 26.04.1 - the engine stopped unexpectedly (exit code -6)

本文

Similar to https://github.com/Niko1221/Strata/issues/579 - short prompt works correctly, long prompt errors out after some thinking:

> the engine stopped unexpectedly (exit code -6)

Fresh Strata install `v0.1.41`, running `./setup.sh` choosing all defaults. Ran via DeepSeek Harness (manually setting context size same as Strata served context size).

`sudo dmesg | grep -i -E 'killed process|out of memory'` prints empty; `sudo dmesg -T | grep -iE "amdgpu|timeout|MES|reset|fault"` prints

> [Thu Oct  8 15:49:38 2026] pid_max: default: 32768 minimum: 301
> [Thu Oct  8 15:49:38 2026] NMI watchdog: Enabled. Permanently consumes one hw-PMU counter.
> [Thu Oct  8 15:49:39 2026] iommu: Default domain type: Translated
> [Thu Oct  8 15:49:39 2026] NetLabel:  unlabeled traffic allowed by default
> [Thu Oct  8 15:49:39 2026] PCI: CLS 64 bytes, default 64
> [Thu Oct  8 15:49:39 2026] workingset: timestamp_bits=36 max_order=23 bucket_order=0
> [Thu Oct  8 15:49:39 2026] Estimated ratio of average max frequency by base frequency (times 1024): 1086
> [Thu Oct  8 15:49:39 2026] x86/amd: Previous system reset reason [0x00080800]: software wrote 0x6 to reset control register 0xCF9
> [Thu Oct  8 15:49:39 2026] systemd[1]: Queued start job for default target initrd.target.
> [Thu Oct  8 15:49:39 2026] systemd-journald[244]: Collecting audit messages is disabled.
> [Thu Oct  8 15:49:39 2026] nvme nvme0: 16/0/0 default/read/poll queues
> [Thu Oct  8 15:50:15 2026] systemd[1]: Reached target cryptsetup.target - Local Encrypted Volumes.
> [Thu Oct  8 15:50:15 2026] systemd[1]: Reached target integritysetup.target - Local Integrity Protected Volumes.
> [Thu Oct  8 15:50:15 2026] systemd[1]: Reached target veritysetup.target - Local Verity Protected Volumes.
> [Thu Oct  8 15:50:15 2026] systemd[1]: Listening on systemd-factory-reset.socket - Factory Reset Management.
> [Thu Oct  8 15:50:15 2026] systemd[1]: Mounting dev-mqueue.mount - POSIX Message Queue File System...
> [Thu Oct  8 15:50:15 2026] systemd[1]: Mounted dev-mqueue.mount - POSIX Message Queue File System.
> [Thu Oct  8 15:50:15 2026] systemd-journald[2384]: Collecting audit messages is disabled.
> [Thu Oct  8 15:50:15 2026] r8169 0000:07:00.0 eth0: jumbo features [frames: 9194 bytes, tx checksumming: ko]
> [Thu Oct  8 15:50:15 2026] r8169 0000:0a:00.0 eth1: jumbo features [frames: 16362 bytes, tx checksumming: ko]
> [Thu Oct  8 15:50:16 2026] amdgpu: Virtual CRAT table created for CPU
> [Thu Oct  8 15:50:16 2026] amdgpu: Topology: Add CPU node
> [Thu Oct  8 15:50:16 2026] amdgpu 0000:03:00.0: enabling device (0006 -> 0007)
> [Thu Oct  8 15:50:16 2026] amdgpu 0000:03:00.0: initializing kernel modesetting (IP DISCOVERY 0x1002:0x7550 0x1EAE:0x8811 0xC0).
> [Thu Oct  8 15:50:16 2026] amdgpu 0000:03:00.0: register mmio base: 0xF6D00000
> [Thu Oct  8 15:50:16 2026] amdgpu 0000:03:00.0: register mmio size: 524288
> [Thu Oct  8 15:50:16 2026] amdgpu 0000:03:00.0: detected ip block number 0 <common_v1_0_0> (soc24_common)
> [Thu Oct  8 15:50:16 2026] amdgpu 0000:03:00.0: detected ip block number 1 <gmc_v12_0_0> (gmc_v12_0)
> [Thu Oct  8 15:50:16 2026] amdgpu 0000:03:00.0: detected ip block number 2 <ih_v7_0_0> (ih_v7_0)
> [Thu Oct  8 15:50:16 2026] amdgpu 0000:03:00.0: detected ip block number 3 <psp_v14_0_0> (psp)
> [Thu Oct  8 15:50:16 2026] amdgpu 0000:03:00.0: detected ip block number 4 <smu_v14_0_0> (smu)
> [Thu Oct  8 15:50:16 2026] amdgpu 0000:03:00.0: detected ip block number 5 <dce_v1_0_0> (dm)
> [Thu Oct  8 15:50:16 2026] amdgpu 0000:03:00.0: detected ip block number 6 <gfx_v12_0_0> (gfx_v12_0)
> [Thu Oct  8 15:50:16 2026] amdgpu 0000:03:00.0: detected ip block number 7 <sdma_v7_0_0> (sdma_v7_0)
> [Thu Oct  8 15:50:16 2026] amdgpu 0000:03:00.0: detected ip block number 8 <vcn_v5_0_0> (vcn_v5_0_0)
> [Thu Oct  8 15:50:16 2026] amdgpu 0000:03:00.0: detected ip block number 9 <jpeg_v5_0_0> (jpeg_v5_0_0)
> [Thu Oct  8 15:50:16 2026] amdgpu 0000:03:00.0: detected ip block number 10 <mes_v12_0_0> (mes_v12_0)
> [Thu Oct  8 15:50:16 2026] amdgpu 0000:03:00.0: Fetched VBIOS from VFCT
> [Thu Oct  8 15:50:16 2026] amdgpu 0000:03:00.0: [drm] ATOM BIOS: 113-EXT108828-100
> [Thu Oct  8 15:50:16 2026] amdgpu 0000:03:00.0: vgaarb: deactivate vga console
> [Thu Oct  8 15:50:16 2026] amdgpu 0000:03:00.0: Trusted Memory Zone (TMZ) feature not supported
> [Thu Oct  8 15:50:16 2026] amdgpu 0000:03:00.0: MEM ECC is not presented.
> [Thu Oct  8 15:50:16 2026] amdgpu 0000:03:00.0: SRAM ECC is not presented.
> [Thu Oct  8 15:50:16 2026] amdgpu 0000:03:00.0: vm size is 262144 GB, 4 levels, block size is 9-bit, fragment size is 9-bit
> [Thu Oct  8 15:50:16 2026] amdgpu 0000:03:00.0: VRAM: 16304M 0x0000008000000000 - 0x00000083FAFFFFFF (16304M used)
> [Thu Oct  8 15:50:16 2026] amdgpu 0000:03:00.0: GART: 512M 0x0000000000000000 - 0x000000001FFFFFFF
> [Thu Oct  8 15:50:16 2026] amdgpu 0000:03:00.0: [drm] Detected VRAM RAM=16304M, BAR=16384M
> [Thu Oct  8 15:50:16 2026] amdgpu 0000:03:00.0: [drm] RAM width 256bits GDDR6
> [Thu Oct  8 15:50:16 2026] amdgpu 0000:03:00.0:  16304M of VRAM memory ready
> [Thu Oct  8 15:50:16 2026] amdgpu 0000:03:00.0:  15576M of GTT memory ready.
> [Thu Oct  8 15:50:16 2026] amdgpu 0000:03:00.0: [drm] GART: num cpu pages 131072, num gpu pages 131072
> [Thu Oct  8 15:50:16 2026] amdgpu 0000:03:00.0: [drm] PCIE GART of 512M enabled (table at 0x00000083DAB00000).
> [Thu Oct  8 15:50:16 2026] amdgpu 0000:03:00.0: [drm] Loading DMUB firmware via PSP: version=0x0A000800
> [Thu Oct  8 15:50:16 2026] amdgpu 0000:03:00.0: [VCN instance 0] Found VCN firmware Version ENC: 1.12 DEC: 9 VEP: 0 Revision: 15
> [Thu Oct  8 15:50:16 2026] amdgpu 0000:03:00.0: MES: vmid_mask_mmhub 0x0000ff00, vmid_mask_gfxhub 0x0000ff00
> [Thu Oct  8 15:50:16 2026] amdgpu 0000:03:00.0: MES: gfx_hqd_mask 0x000000fe, compute_hqd_mask 0x0000000c, sdma_hqd_mask 0x000000fc
> [Thu Oct  8 15:50:17 2026] amdgpu 0000:03:00.0: RAP: optional rap ta ucode is not available
> [Thu Oct  8 15:50:17 2026] amdgpu 0000:03:00.0: SECUREDISPLAY: optional securedisplay ta ucode is not available
> [Thu Oct  8 15:50:17 2026] amdgpu 0000:03:00.0: smu driver if version = 0x0000002e, smu fw if version = 0x00000033, smu fw program = 0, smu fw version = 0x00684c00 (104.76.0)
> [Thu Oct  8 15:50:17 2026] amdgpu 0000:03:00.0: SMU is initialized successfully!
> [Thu Oct  8 15:50:17 2026] amdgpu 0000:03:00.0: [drm] Display Core v3.2.369 initialized on DCN 4.0.1
> [Thu Oct  8 15:50:17 2026] amdgpu 0000:03:00.0: [drm] DP-HDMI FRL PCON supported
> [Thu Oct  8 15:50:17 2026] amdgpu 0000:03:00.0: [drm] DMUB hardware initialized: version=0x0A000800
> [Thu Oct  8 15:50:17 2026] snd_hda_intel 0000:03:00.1: bound 0000:03:00.0 (ops amdgpu_dm_audio_component_bind_ops [amdgpu])
> [Thu Oct  8 15:50:17 2026] amdgpu 0000:03:00.0: [drm] DP-1: PSR support 0, DC PSR ver -1, sink PSR ver 0 DPCD caps 0x0 su_y_granularity 0
> [Thu Oct  8 15:50:17 2026] amdgpu 0000:03:00.0: [drm] DP-2: PSR support 0, DC PSR ver -1, sink PSR ver 0 DPCD caps 0x0 su_y_granularity 0
> [Thu Oct  8 15:50:17 2026] amdgpu 0000:03:00.0: [drm] DP-3: PSR support 0, DC PSR ver -1, sink PSR ver 0 DPCD caps 0x0 su_y_granularity 0
> [Thu Oct  8 15:50:17 2026] amdgpu 0000:03:00.0: [drm] HDMI-A-1: PSR support 0, DC PSR ver -1, sink PSR ver 0 DPCD caps 0x0 su_y_granularity 0
> [Thu Oct  8 15:50:17 2026] amdgpu 0000:03:00.0: program CP_MES_CNTL : 0x4000000
> [Thu Oct  8 15:50:17 2026] amdgpu 0000:03:00.0: program CP_MES_CNTL : 0xc000000
> [Thu Oct  8 15:50:17 2026] amdgpu: Virtual CRAT table created for GPU
> [Thu Oct  8 15:50:17 2026] amdgpu: Topology: Add GPU node [0x1002:0x7550]
> [Thu Oct  8 15:50:17 2026] amdgpu 0000:03:00.0: SE 4, SH per SE 2, CU per SH 8, active_cu_number 64
> [Thu Oct  8 15:50:17 2026] amdgpu 0000:03:00.0: ring gfx_0.0.0 uses VM inv eng 0 on hub 0
> [Thu Oct  8 15:50:17 2026] amdgpu 0000:03:00.0: ring comp_1.0.0 uses VM inv eng 1 on hub 0
> [Thu Oct  8 15:50:17 2026] amdgpu 0000:03:00.0: ring comp_1.1.0 uses VM inv eng 4 on hub 0
> [Thu Oct  8 15:50:17 2026] amdgpu 0000:03:00.0: ring comp_1.0.1 uses VM inv eng 7 on hub 0
> [Thu Oct  8 15:50:17 2026] amdgpu 0000:03:00.0: ring comp_1.1.1 uses VM inv eng 8 on hub 0
> [Thu Oct  8 15:50:17 2026] amdgpu 0000:03:00.0: ring sdma0 uses VM inv eng 9 on hub 0
> [Thu Oct  8 15:50:17 2026] amdgpu 0000:03:00.0: ring sdma1 uses VM inv eng 10 on hub 0
> [Thu Oct  8 15:50:17 2026] amdgpu 0000:03:00.0: ring vcn_unified_0 uses VM inv eng 0 on hub 8
> [Thu Oct  8 15:50:17 2026] amdgpu 0000:03:00.0: ring jpeg_dec uses VM inv eng 1 on hub 8
> [Thu Oct  8 15:50:17 2026] amdgpu: HMM registered 16304MB device memory
> [Thu Oct  8 15:50:17 2026] amdgpu 0000:03:00.0: Using BACO for runtime pm
> [Thu Oct  8 15:50:17 2026] amdgpu 0000:03:00.0: [drm] Registered 4 planes with drm panic
> [Thu Oct  8 15:50:17 2026] [drm] Initialized amdgpu 3.64.0 for 0000:03:00.0 on minor 1
> [Thu Oct  8 15:50:17 2026] fbcon: amdgpudrmfb (fb0) is primary device
> [Thu Oct  8 15:50:17 2026] amdgpu 0000:03:00.0: [drm] fb0: amdgpudrmfb frame buffer device
> [Thu Oct  8 15:53:45 2026] workqueue: svm_range_restore_work [amdgpu] hogged CPU for >10000us 4 times, consider switching to WQ_UNBOUND
> [Thu Oct  8 15:54:44 2026] workqueue: svm_range_restore_work [amdgpu] hogged CPU for >10000us 5 times, consider switching to WQ_UNBOUND
> [Thu Oct  8 15:54:46 2026] audit: type=1400 audit(1791467686.653:254): apparmor="AUDIT" operation="userns_create" class="namespace" info="Userns create - transitioning profile" profile="unconfined" pid=9548 comm="(ubuntu-insigh)" requested="userns_create" target="unprivileged_userns" execpath="/usr/lib/systemd/systemd-executor"
> [Thu Oct  8 15:54:47 2026] workqueue: svm_range_restore_work [amdgpu] hogged CPU for >10000us 7 times, consider switching to WQ_UNBOUND
> [Thu Oct  8 15:55:56 2026] workqueue: amdgpu_gfx_profile_idle_work_handler [amdgpu] hogged CPU for >10000us 4 times, consider switching to WQ_UNBOUND
> [Thu Oct  8 15:57:46 2026] audit: type=1400 audit(1791467866.621:257): apparmor="AUDIT" operation="userns_create" class="namespace" info="Userns create - transitioning profile" profile="unconfined" pid=10313 comm="(ubuntu-insigh)" requested="userns_create" target="unprivileged_userns" execpath="/usr/lib/systemd/systemd-executor"

My Strata log: [strata-q2_0.log](https://github.com/user-attachments/files/33207901/strata-q2_0.log)

`amd-smi` after Strata errored:
> +------------------------------------------------------------------------------+
> | AMD-SMI            27.1.0+7da43026                                           |
> | OS kernel Version: 7.0.0-38-generic                                          |
> | ROCm Version:      10.1.0                                                    |
> | VBIOS Version:     00131164                                                  |
> | Platform:          Linux Baremetal                                           |
> |-------------------------------------+----------------------------------------|
> | BDF                        GPU-Name | Mem-Uti   Temp   UEC       Power-Usage |
> | GPU  HIP-ID  OAM-ID  Partition-Mode | GFX-Uti    Fan               Mem-Usage |
> |=====================================+========================================|
> | 0000:03:00.0  AMD Radeon RX 9070 XT | 0 %      32 °C   0            10/340 W |
> |   0       0     N/A             N/A | 4 %      0.0 %            572/16304 MB |
> +-------------------------------------+----------------------------------------+
> +------------------------------------------------------------------------------+
> | Processes:                                                                   |
> |  GPU      PID  Process Name     GTT_MEM  VRAM_MEM  MEM_USAGE   CU %     SDMA |
> |==============================================================================|
> |  No running processes found                                                  |
> +---------------------------------------------------------------------------