AdaptiveCpp
CTranslate2
AdaptiveCpp | CTranslate2 | |
---|---|---|
19 | 14 | |
1,042 | 2,825 | |
2.4% | 4.7% | |
9.7 | 8.9 | |
9 days ago | 9 days ago | |
C++ | C++ | |
BSD 2-clause "Simplified" License | MIT License |
Stars - the number of stars that a project has on GitHub. Growth - month over month growth in stars.
Activity is a relative number indicating how actively a project is being developed. Recent commits have higher weight than older ones.
For example, an activity of 9.0 indicates that a project is amongst the top 10% of the most actively developed projects that we are tracking.
AdaptiveCpp
-
What Every Developer Should Know About GPU Computing
Sapphire Rapids is a CPU.
AMD's primary focus for a GPU software ecosystem these days seems to be implementing CUDA with s/cuda/hip, so AMD directly supports and encourages running GPU software written in CUDA on AMD GPUs.
The only implementation for sycl on AMD GPUs that I can find is a hobby project that apparently is not allowed to use either the 'hip' or 'sycl' names. https://github.com/AdaptiveCpp/AdaptiveCpp
-
AMD May Get Across the CUDA Moat
Not natively, but AdaptiveCpp (previously hiSycl, then OpenSycl) has a single source single compiler pass, where they basically store LLVM IR as an intermediate representation.
https://github.com/AdaptiveCpp/AdaptiveCpp/blob/develop/doc/...
Performance penalty was within ew precents, at least according to the paper (figure 9 and 10)
-
Offloading standard C++ PSTL to Intel, NVIDIA and AMD GPUs with AdaptiveCpp
AdaptiveCpp (formerly known as hipSYCL) is an independent, open source, clang-based heterogeneous C++ compiler project. I thought some of you might be interested in knowing that we recently added support to offload standard C++ parallel STL algorithms to GPUs from all major vendors. E.g.:
-
AMD's HIPRT Working Its Way To Blender With ~25% Faster Rendering
In fact SYCL was initially called hipSYCL because it is based on AMD's ROCm/HIP. AMD had hipSYCL code running on the Frontier supercomputer four years ago at least and continues to support it.
-
hipSYCL can now generate a binary that runs on any Intel/NVIDIA/AMD GPU - in a single compiler pass. It is now the first single-pass SYCL compiler, and the first with unified code representation across backends.
Apple Silicon support through Metal is something that is actively discussed in hipSYCL. See https://github.com/illuhad/hipSYCL/issues/864 https://github.com/illuhad/hipSYCL/issues/460 (loooong discussion)
-
Bringing Nvidia® and AMD support to oneAPI
But really, the DPC++ part of oneAPI (which is many APIs) is just SYCL + extensions, and there are several other SYCL implementations which have already featured CUDA and Hip (AMD) support for a long time. The most popular and widely-used is hipSYCL, which we've been using in an HPC context on NV hardware for over 4 years now.
-
Intel oneAPI 2023 Released - AMD & NVIDIA Plugins Available
Unfortunately, the AMD and Nvidia plugins are proprietary. AMD users are probably better served with hipSYCL, if they somehow find an application using SYCL...
-
There is framework for everything.
Also, you might want to take a look at an implementation like hipSYCL :)
-
The Next Platform: "Intel Takes The SYCL To Nvidia's CUDA With Migration Tool"
Yup. SYCL is the future: https://github.com/illuhad/hipSYCL
-
Phoronix: "Intel's Vulkan Linux Driver Adds Experimental Mesh Shader Support For DG2/Alchemist"
ROCm is completely independent from these. It's a compute stack containing an OpenCL implementation for Radeon GPUs, plus a CUDA-like language called HIP which can be compiled to either device code for Radeon GPUs or to PTX to work with Nvidia GPUs. However, some researchers also created hipSYCL that allows SYCL to run atop HIP; you can think of it like DXVK - the program contains the DirectX/SYCL API, and DXVK/hipSYCL converts it to Vulkan/HIP (with one difference - DXVK does the conversion at runtime, while hipSYCL does it at compile time).
CTranslate2
- Creando Subtítulos Automáticos para Vídeos con Python, Faster-Whisper, FFmpeg, Streamlit, Pillow
-
Distil-Whisper: distilled version of Whisper that is 6 times faster, 49% smaller
Just a point of clarification - faster-whisper references it but ctranslate2[0] is what's really doing the magic here.
Ctranslate2 is a sleeper powerhouse project that enables a lot. They should be up front and center and get the credit they deserve.
[0] - https://github.com/OpenNMT/CTranslate2
-
A Raspberry Pi 5 is better than two Pi 4S
We'd love to move beyond Nvidia.
The issue (among others) is we achieve the speech recognition performance we do largely thanks to ctranslate2[0]. They've gone on the record saying that they essentially have no interest in ROCm[1].
Of course with open source anything is possible but we see this as being one of several fundamental issues in supporting AMD GPGPU hardware.
[0] - https://github.com/OpenNMT/CTranslate2
[1] - https://github.com/OpenNMT/CTranslate2/issues/1072
-
AMD May Get Across the CUDA Moat
> While I agree that it's much more effort to get things working on AMD cards than it is with Nvidia, I was a bit surprised to see this comment mention Whisper being an example of "5-10x as performant".
It easily is. See the benchmarks[0] from faster-whisper which uses Ctranslate2. That's 5x faster than OpenAI reference code on a Tesla V100. Needless to say something like a 4080 easily multiplies that.
> https://www.tomshardware.com/news/whisper-audio-transcriptio... is a good example of Nvidia having no excuses being double the price when it comes to Whisper inference, with 7900XTX being directly comparable with 4080, albeit with higher power draw. To be fair it's not using ROCm but Direct3D 11, but for performance/price arguments sake that detail is not relevant.
With all due respect to the author of the article this is "my first entry into ML" territory. They talk about a 5-10 second delay, my project can do sub 1 second times[1] even with ancient GPUs thanks to Ctranslate2. I don't have an RTX 4080 but if you look at the performance stats for the closest thing (RTX 4090) the performance numbers are positively bonkers - completely untouchable for anything ROCm based. Same goes for the other projects I linked, lmdeploy does over 100 tokens/s in a single session with LLama2 13b on my RTX 4090 and almost 600 tokens/s across eight simultaneous sessions.
> EDIT: Also using CTranslate2 as an example is not great as it's actually a good showcase why ROCm is so far behind CUDA: It's all about adapting the tech and getting the popular libraries to support it. Things usually get implemented in CUDA first and then would need additional effort to add ROCm support that projects with low amount of (possibly hobbyist) maintainers might not have available. There's even an issue in CTranslate2 where they clearly state no-one is working to get ROCm supported in the library. ( https://github.com/OpenNMT/CTranslate2/issues/1072#issuecomm... )
I don't understand what you're saying here. It (along with the other projects I linked) are fantastic examples of just how far behind the ROCm ecosystem is. ROCm isn't even on the radar for most of them as your linked issue highlights.
Things always get implemented in CUDA first (ten years in this space and I've never seen ROCm first) and ROCm users either wait months (minimum) for sub-par performance or never get it at all.
[0] - https://github.com/guillaumekln/faster-whisper#benchmark
[1] - https://heywillow.io/components/willow-inference-server/#ben...
-
StreamingLLM: Efficient streaming technique enable infinite sequence lengths
Etc.
Now, what this allows you to do is reuse the attention computed from the previous turns (since the prefix is the same).
In practice, people often have a system prompt before the conversation history, which (as far a I can tell) makes this technique not applicable (the input prefix will change as soon as the conversation history is long enough that we need to start dropping the oldest turns).
In such case, what you could do is to cache at least the system prompt. This is also possible with https://github.com/OpenNMT/CTranslate2/blob/2203ad5c8baf878a...
-
Faster Whisper Transcription with CTranslate2
The original Whisper implementation from OpenAI uses the PyTorch deep learning framework. On the other hand, faster-whisper is implemented using CTranslate2 [1] which is a custom inference engine for Transformer models. So basically it is running the same model but using another backend, which is specifically optimized for inference workloads.
[1] https://github.com/OpenNMT/CTranslate2
-
Explore large language models on any computer with 512MB of RAM
FLAN-T5 models generally perform well for their size, but they are encode-decoder models, and they aren't as widely supported for efficient inference. I wanted students to be able to run everything locally on CPU, so I was ideally hoping for something that supported quantization for CPU inference. I explored llama.cpp and GGML, but ultimately landed on ctranslate2 for inference.
- CTranslate2: An efficient inference engine for Transformer models
-
[D] Faster Flan-T5 inference
You can also check out the CTranslate2 library which supports efficient inference of T5 models, including 8-bit quantization on CPU and GPU. There is a usage example in the documentation.
- Running large language models like ChatGPT on a single GPU
What are some alternatives?
ROCm - AMD ROCm™ Software - GitHub Home [Moved to: https://github.com/ROCm/ROCm]
vllm - A high-throughput and memory-efficient inference and serving engine for LLMs
HIP-CPU - An implementation of HIP that works on CPUs, across OSes.
sentencepiece - Unsupervised text tokenizer for Neural Network-based text generation.
triSYCL - Generic system-wide modern C++ for heterogeneous platforms with SYCL from Khronos Group
FlexGen - Running large language models like OPT-175B/GPT-3 on a single GPU. Focusing on high-throughput generation. [Moved to: https://github.com/FMInference/FlexGen]
HIP - HIP: C++ Heterogeneous-Compute Interface for Portability
OpenNMT-Tutorial - Neural Machine Translation (NMT) tutorial. Data preprocessing, model training, evaluation, and deployment.
cuda-api-wrappers - Thin C++-flavored header-only wrappers for core CUDA APIs: Runtime, Driver, NVRTC, NVTX.
oneDNN - oneAPI Deep Neural Network Library (oneDNN)
cuda_memtest - Fork of CUDA GPU memtest :eyeglasses:
faster-whisper - Faster Whisper transcription with CTranslate2