Deion of the problem Hi, I'm trying to ru

First, you should use the new Gramine repo: <a href="https://github.com/gramineproject

<a class="user-mention notranslate" data-hovercard-type="user" data-hovercard-url="/us

pytorch sample config for better performance about graphene HOT 2 CLOSED

salsal97 commented on July 3, 2024

pytorch sample config for better performance

from graphene.

Comments (2)

dimakuv commented on July 3, 2024

First, you should use the new Gramine repo: https://github.com/gramineproject/gramine (Graphene was renamed to Gramine and moved under a new GitHub org). This "Graphene" repo is outdated.

Second, what is your machine? How much EPC does it have? If you have an older, wimpy SGX-enabled machine with only 96MB - 128MB of EPC, you will see pretty bad performance on big workloads like PyTorch. Please also read through this: https://gramine.readthedocs.io/en/latest/manifest-syntax.html#enclave-size and this: https://gramine.readthedocs.io/en/latest/devel/performance.html#choice-of-sgx-machine

from graphene.

salsal97 commented on July 3, 2024

@dimakuv
Hi, I used graphene since gramine does not have a pytorch sample as of yet. I'm using an Icelake machine, with 24 cores -

$ is-sgx-available 
SGX supported by CPU: true
SGX1 (ECREATE, EENTER, ...): true
SGX2 (EAUG, EACCEPT, EMODPR, ...): true
Flexible Launch Control (IA32_SGXPUBKEYHASH{0..3} MSRs): true
SGX extensions for virtualizers (EINCVIRTCHILD, EDECVIRTCHILD, ESETCONTEXT): false
Extensions for concurrent memory management (ETRACKC, ELDBC, ELDUC, ERDINFO): false
CET enclave attributes support (See Table 37-5 in the SDM): false
Key separation and sharing (KSS) support (CONFIGID, CONFIGSVN, ISVEXTPRODID, ISVFAMILYID report fields): true
Max enclave size (32-bit): 0x80000000
Max enclave size (64-bit): 0x100000000000000
EPC size: 0x200000000
SGX driver loaded: true
AESMD installed: true
SGX PSW/libsgx installed: true

from graphene.

Recommend Projects