Google Cloud tesla K80, only one device showing up?
I've setup an instance on Google Cloud with the following specs: 4 vCPUs, 15 GB memory, 1 Tesla K80 GPU
Tesla K80 consists of 2 GPU units and each should show up as a separate device in the nvidia's logs. However, when I run nvidia-smi in the shell it shows only one.
Image:
Does anyone know how to solve this issue? Is this because my cloud GPU quota is one and hence only one device is being used?
Additional logs:
==============NVSMI LOG==============
Timestamp : Tue Mar 13 16:05:42 2018
Driver Version : 390.30
Attached GPUs : 1
GPU 00000000:00:04.0
Product Name : Tesla K80
Product Brand : Tesla
Display Mode : Disabled
Display Active : Disabled
Persistence Mode : Disabled
Accounting Mode : Disabled
Accounting Mode Buffer Size : 1920
Driver Model
Current : N/A
Pending : N/A
Serial Number : 0320717075175
GPU UUID : GPU-a3a146ad-aed1-d5ef-1e76-2565c1e20a13
Minor Number : 0
VBIOS Version : 80.21.25.00.01
MultiGPU Board : No
Board ID : 0x4
GPU Part Number : 900-22080-6300-001
Inforom Version
Image Version : 2080.0200.00.04
OEM Object : 1.1
ECC Object : 3.0
Power Management Object : N/A
GPU Operation Mode
Current : N/A
Pending : N/A
GPU Virtualization Mode
Virtualization mode : Pass-Through
PCI
Bus : 0x00
Device : 0x04
Domain : 0x0000
Device Id : 0x102D10DE
Bus Id : 00000000:00:04.0
Sub System Id : 0x106C10DE
GPU Link Info
PCIe Generation
Max : 3
Current : 3
Link Width
Max : 16x
Current : 16x
Bridge Chip
Type : N/A
Firmware : N/A
Replays since reset : 0
Tx Throughput : N/A
Rx Throughput : N/A
Fan Speed : N/A
Performance State : P0
Clocks Throttle Reasons
Idle : Not Active
Applications Clocks Setting : Active
SW Power Cap : Not Active
HW Slowdown : Not Active
HW Thermal Slowdown : N/A
HW Power Brake Slowdown : N/A
Sync Boost : Not Active
SW Thermal Slowdown : Not Active
Display Clock Setting : Not Active
FB Memory Usage
Total : 11441 MiB
Used : 10930 MiB
Free : 511 MiB
BAR1 Memory Usage
Total : 16384 MiB
Used : 3 MiB
Free : 16381 MiB
Compute Mode : Default
Utilization
Gpu : 54 %
Memory : 1 %
Encoder : 0 %
Decoder : 0 %
Encoder Stats
Active Sessions : 0
Average FPS : 0
Average Latency : 0
Ecc Mode
Current : Enabled
Pending : Enabled
ECC Errors
Volatile
Single Bit
Device Memory : 0
Register File : 0
L1 Cache : 0
L2 Cache : 0
Texture Memory : 0
Texture Shared : N/A
CBU : N/A
Total : 0
Double Bit
Device Memory : 0
Register File : 0
L1 Cache : 0
L2 Cache : 0
Texture Memory : 0
Texture Shared : N/A
CBU : N/A
Total : 0
Aggregate
Single Bit
Device Memory : 0
Register File : 0
L1 Cache : 0
L2 Cache : 0
Texture Memory : 0
Texture Shared : N/A
CBU : N/A
Total : 0
Double Bit
Device Memory : 7
Register File : 0
L1 Cache : 0
L2 Cache : 0
Texture Memory : 0
Texture Shared : N/A
CBU : N/A
Total : 7
Retired Pages
Single Bit ECC : 0
Double Bit ECC : 1
Pending : No
Temperature
GPU Current Temp : 56 C
GPU Shutdown Temp : 93 C
GPU Slowdown Temp : 88 C
GPU Max Operating Temp : N/A
Memory Current Temp : N/A
Memory Max Operating Temp : N/A
Power Readings
Power Management : Supported
Power Draw : 66.18 W
Power Limit : 149.00 W
Default Power Limit : 149.00 W
Enforced Power Limit : 149.00 W
Min Power Limit : 100.00 W
Max Power Limit : 175.00 W
Clocks
Graphics : 562 MHz
SM : 562 MHz
Memory : 2505 MHz
Video : 540 MHz
Applications Clocks
Graphics : 562 MHz
Memory : 2505 MHz
Default Applications Clocks
Graphics : 562 MHz
Memory : 2505 MHz
Max Clocks
Graphics : 875 MHz
SM : 875 MHz
Memory : 2505 MHz
Video : 540 MHz
Max Customer Boost Clocks
Graphics : N/A
Clock Policy
Auto Boost : On
Auto Boost Default : On
Solution
Digging through the cloud platform website, I found an answer to this. Instead of removing the question, I am leaving this here, in case someone else faces the same issue.
The Tesla K80 is '1 GPU board' consisting of 2 GPUs. If the number of GPUs on your VM instance on Google Cloud Platform is set to 1, you'd be allocated 1 GPU or half of a board.
- oauth 2 parameters can only have a single value: scope
- Correct use of gcloud --sort-by combined with --limit
- App attestation failed with Firebase App check in release on Android
- Why doesn't the offline cloud firestore documentation have code for offline query indexes for flutter?
- _CHANGE_TYPE not working when streaming Google Big Query rows from Pub/Sub
- Migrating from legacy Google FCM to HTTP v1 for push notifications
- Google OAuth 2 Error 400: redirect_uri_mismatch but redirect uri is compliant and already registered in Google Cloud Console
- Can we update a field due to its value without reading it in firebase?
- Use process.env variables inside next.config.js
- What is the best practice for decoding Firestore documents in a listener for Swift 5?
- How to install Google Cloud SDK app-engine-python on Debian 12?
- Bigquery service account restricted to a dataset
- Gcloud Pub/Sub: Total timeout of API google.pubsub.v1.Publisher exceeded 60000 milliseconds before any response was received
- Keep getting the error Google Sign-In Error: PlatformException(sign_in_failed, com.google.android.gms.common.api.ApiException: 10: , null, null)
- Error 400 invalid JSON Payload Unknown name generationConfig on an AI API Curl command
- How to access cloud run environment variables in Dockerfile
- Error 400: Invalid JSON payload received when calling Google Generative Language API
- GCP Bucket Permissions by Prefix
- database connection URI string for GCP cloudsql postgres database?
- Who control Google Cloud platform filtering?
- gcloud cli command create eventarc errors
- Is there a way to block HTTP on GCP AppEngine
- How to generate GOOGLE_CLIENT_ID and GOOGLE_CLIENT_SECRET without a domain?
- How to change the source GitHub repository for a deployed project on Google Cloud Run?
- Not able to access Nexus console on my Ubuntu VM instance in GCP
- VertexAI Tabular AutoML rejecting rows containing nulls
- How to access Google Cloud Speech-to-text v2 API through HTTP/REST
- Importing a Model with Scikit Learn on Vertex
- Is there a simple way to host a JSON document you can read and update in Google Cloud Platform?
- how can I get my gcloud user creds into a container securely and use them to impersonate a service account when testing locally?