Repository navigation
VQE performance on the CUDA-Q #1684
Replies: 1 comment
|
This is a very good question—and the short answer is:
But there is meaningful context on where CUDA-Q sits in the landscape. 1. First: what CUDA-Q VQE is actually optimized forCUDA-Q VQE (hybrid workflow) is primarily designed to accelerate:
So it improves:
2. Why “comparison with quantum computers” is trickyComparing CUDA-Q VQE vs real QPUs is not straightforward because: Different objectives:
So CUDA-Q is:
3. CUDA-Q vs classical solvers (important nuance)There are partial comparisons in NVIDIA materials and community experiments: CUDA-Q tends to outperform when:✔ large statevector workloads
✔ repeated Hamiltonian expectation evaluation
Classical solvers still win when:✔ small / structured problems
4. Why CUDA-Q VQE is fast (core reason)CUDA-Q gains come from: GPU acceleration of:
So instead of: you get: 5. Why there is no “standard benchmark paper”There are three reasons: (A) VQE is not a single algorithmIt depends heavily on:
So results vary wildly. (B) CUDA-Q evolves quicklyBenchmarks become outdated fast due to:
(C) QPU performance is not stableReal hardware varies by:
So comparisons are not reproducible. 6. What does exist in practiceFrom NVIDIA CUDA-Q ecosystem + papers/blogs: ✔ GPU vs CPU VQE comparisons:
✔ Hybrid workflows:
But:
7. Key conceptual takeawayThink of CUDA-Q VQE as:
not as:
8. When CUDA-Q VQE is most useful✔ Prototyping ansatz circuits 9. Direct answer
If you wantI can also break down:
If this answer helped or pointed you in the right direction, I'd appreciate it if you could mark it as the accepted answer so it's easier for others with the same issue to find. Also, if you found my contribution useful, I'd appreciate it if you could check out my GitHub profile, follow me, and star any repositories you find interesting. GitHub: https://github.com/Advait251206 |
Uh oh!
There was an error while loading. Please reload this page.
Hi,
I'm wondering if there has been any study on comparing VQE on hybrid CUDA-Q compared with the current quantum computers or classical solvers, I would appreciate if anyone can help. I see that there is a recent article on VQE (https://developer.nvidia.com/blog/new-nvidia-cuda-q-features-boost-quantum-application-performance/), but it seems to only include different GPU simulations, no comparison with other computational unit...
Thanks
Ami
All reactions