Skip to content

Commit c525bb7

Browse files
Merge pull request #31 from codeplaysoftware/add-codasip-update-posts
Add Codasip Update Posts
2 parents 921fbfb + acd26c6 commit c525bb7

11 files changed

+196
-0
lines changed
Lines changed: 18 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,18 @@
1+
---
2+
title: "RISC-V customization, HW/SW co-optimization, and custom compute"
3+
date: 2023-05-02
4+
layout: update
5+
tags:
6+
- codasip
7+
- risc-v
8+
- customization
9+
link: https://codasip.com/2023/05/02/riscv-customization-hardware-software-co-optimization-custom-compute/
10+
---
11+
12+
Do we still need to introduce and define RISC-V? You know, the open-source instruction set architecture (ISA) that is
13+
gaining popularity thanks to its flexibility, scalability, and modularity. Okay, we just did, just to be sure we are all
14+
on the same page. One of the key benefits and the main “raison d’être” of RISC-V is the possibility to tailor both the
15+
instruction set (ISA) and the internal design (microarchitecture) of the processor to meet specific application
16+
requirements. This customization capability extends to custom compute solutions, enabling developers to create hardware
17+
optimized for their workloads. In this blog post, let’s explore the benefits of RISC-V customization and custom compute,
18+
and industry applications.
Lines changed: 18 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,18 @@
1+
---
2+
title: "No one-size-fits-all approach to RISC-V processor optimization"
3+
date: 2023-05-23
4+
layout: update
5+
tags:
6+
- codasip
7+
- risc-v
8+
- optimization
9+
link: https://codasip.com/2023/05/23/no-one-size-fits-all-approach-to-processor-optimization/
10+
---
11+
12+
As the demand for high-performance processors continues to grow and semiconductor scaling laws continue to show their
13+
limits, the need for processor optimization is inevitable. As I explained in a previous blog, RISC-V is designed to
14+
enable this. However, there is no one-size-fits-all approach to processor optimization. As each workload and each
15+
application will have their own requirements, there are actually different ways to optimize. You can modify a processor
16+
IP at different levels, each with its own benefits. In this blog post, let’s define and explore the different levels of
17+
processor optimization. From configuration to customization, let’s see how you can use them to create optimized
18+
processors that meet specific requirements.
Lines changed: 20 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,20 @@
1+
---
2+
title: "Re-targetable LLVM C/C++ compiler for RISC-V"
3+
date: 2023-07-25
4+
layout: update
5+
tags:
6+
- codasip
7+
- risc-v
8+
- llvm
9+
link: https://codasip.com/2023/07/25/re-targetable-llvm-c-c-plus-plus-compiler-for-riscv/
10+
---
11+
12+
RISC-V is a modular instruction set architecture (ISA) with great customization capabilities that enable innovation and
13+
differentiation without fragmentation. On top of the baseline modules from ratified/standard ISA extensions, such as
14+
integer instructions or floating-point instructions, designers can add custom instructions: pure design freedom! And the
15+
reasons for adding instructions are many: better performance, smaller memory footprint, lower power consumption, or
16+
anything in between.
17+
18+
This means one important thing: the software (the final application(s)) is compiled for the particular RISC-V ISA. The
19+
software development kit (SDK) must know which ISA modules the RISC-V processor implements, so it can automatically
20+
leverage them. This includes both standard instructions and custom instructions.
Lines changed: 20 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,20 @@
1+
---
2+
title: "How the SYCLOPS project democratizes AI acceleration"
3+
date: 2023-08-31
4+
layout: update
5+
tags:
6+
- codasip
7+
- risc-v
8+
- ai
9+
- ml
10+
link: https://codasip.com/2023/08/31/how-the-syclops-project-democratizes-ai-acceleration/
11+
---
12+
13+
Codasip Labs is all about innovation, and specifically the commercialization of that innovation. Naturally, with the
14+
rise of Artificial Intelligence (AI) and Machine Learning (ML), these areas have become a key focus for us. At the
15+
beginning of 2023, we joined the New Horizon Europe Project SYCLOPS (Scaling extreme analYtics with Cross-architecture
16+
acceLeration based on OPen Standards). This project aims to advance AI and data mining for extremely large and diverse
17+
data. The program brings together eight leading European organizations with the goal of achieving ground-breaking
18+
advances in the scalability of extreme data analytics via fully open AI acceleration. Codasip’s role is to develop
19+
domain-specific accelerators based on the RISC-V Vector extension (RVV). We will do this using the design automation
20+
tool Codasip Studio and the CodAL processor description language.
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,16 @@
1+
---
2+
title: "Codasip 700 RISC-V processor family: Bringing the world of Custom Compute to everyone"
3+
date: 2023-10-17
4+
layout: update
5+
tags:
6+
- codasip
7+
- risc-v
8+
link: https://codasip.com/2023/10/17/codasip-700-riscv-processor-family-bringing-the-world-of-custom-compute-to-everyone/
9+
---
10+
11+
Today, technology innovators must have new ways to create differentiated products. How are they supposed to meet the
12+
demand for more computational performance when semiconductor scaling laws are showing their limits? There is only one
13+
way: having a compute that is custom for specific needs. And what do we need for that? Several aspects: Architecture
14+
optimization, application profiling, hardware/software co-optimization, and domain-specific acceleration built on a
15+
strong design foundation. And this is great, but the design process must be as streamlined as possible to improve
16+
efficiency and reduce time to market while allowing companies to take ownership and remain flexible.
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,17 @@
1+
---
2+
title: "Effectively hiding sensitive data with RISC-V Zk and custom instructions"
3+
date: 2024-01-31
4+
layout: update
5+
tags:
6+
- codasip
7+
- risc-v
8+
- cryptographic
9+
link: https://codasip.com/2024/01/31/effectively-hiding-sensitive-data-with-risc-v-zk-and-custom-instructions/
10+
---
11+
12+
Cryptographic hash functions play a critical role in computer security providing a one-way transformation of sensitive
13+
data. Many information-security applications benefit from using hash functions, specifically digital signatures, message
14+
authentication codes, and other forms of authentication. The calculation of hash functions such as SHA512, SHA256, MD5
15+
etc is a potential playground for Custom Compute. This is where the ISA flexibility enabled by RISC-V and empowered by
16+
the Zk extension, as well as the ability to merge inherently sequential bit manipulations in custom instructions help to
17+
improve the performance.
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,21 @@
1+
---
2+
title: "A custom RISC-V vector instruction to accelerate structured-sparse matrix multiplications"
3+
date: 2024-03-20
4+
layout: update
5+
tags:
6+
- codasip
7+
- risc-v
8+
- ai
9+
- matrix
10+
link: https://codasip.com/2024/03/20/a-custom-risc-v-vector-instruction/
11+
---
12+
13+
A novel AI-acceleration paper presents a method to optimize sparse matrix multiplication for machine learning models,
14+
particularly focusing on structured sparsity. Structured sparsity involves a predefined pattern of zero values in the
15+
matrix, unlike unstructured sparsity where zeros can occur anywhere. The research was conducted by Democritus University
16+
of Thrace (DUTH) in Greece and was sponsored by Codasip University Program.
17+
18+
Structured sparsity has emerged as a promising approach to streamline the complexity of modern Machine Learning (ML)
19+
applications and facilitate the handling of sparse data in hardware. Accelerating ML models, whether for training or
20+
inference, heavily relies on efficient execution of equivalent matrix multiplications, which are often performed on
21+
vector processors or custom matrix engines.
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,17 @@
1+
---
2+
title: "Custom Compute for Edge AI: Accelerating innovation with Lund University and Codasip University Program"
3+
date: 2024-04-04
4+
layout: update
5+
tags:
6+
- codasip
7+
- risc-v
8+
- edge
9+
- hpc
10+
- ai
11+
link: https://codasip.com/2024/04/04/custom-compute-for-edge-ai/
12+
---
13+
14+
In recent years, the rapid advancement and adoption of Artificial Intelligence (AI) on the edge has brought about a
15+
surge in development. As AI models like ChatGPT become more prevalent and accurate, the computational requirements for
16+
inference also escalate. This necessitates architectural innovations aimed at reducing both power consumption and
17+
latency.
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,20 @@
1+
---
2+
title: "Sunny skies and electric energy: RISC-V Summit Europe 2024 shines in Munich"
3+
date: 2024-06-28
4+
layout: update
5+
tags:
6+
- codasip
7+
- risc-v
8+
- energy
9+
link: https://codasip.com/2024/06/28/sunny-skies-and-electric-energy-risc-v-summit-europe-2024-shines-in-munich/
10+
---
11+
12+
As I sat on the plane in Boston it’s fair to say that I was curious about what DAC 2024 would bring. The previous year
13+
was much better than I expected but a cold June in San Francisco wasn’t exactly what I was dreaming about. Afterall,
14+
while I was heading to San Francisco with a bunch of other Codasippers some of the Codasip team was headed for the
15+
RISC-V Summit in Munich.
16+
17+
The event is nowhere on the scale that it used to be. In fact, the entire exhibit could probably now fill one floor of
18+
the Moscone. But the level of foot traffic remained high throughout the show and the team and I spoke to a fair number
19+
of people over the 3 days of exhibits. I was impressed with the number of fresh faces and new startups. There was
20+
something interesting to see in all corners of the tradeshow floor. And… incredibly the weather was, well, incredible.
Lines changed: 19 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,19 @@
1+
---
2+
title: "DAC 2024 – Showcasing the future of RISC-V through EDA"
3+
date: 2024-07-16
4+
layout: update
5+
tags:
6+
- codasip
7+
- risc-v
8+
- conference
9+
- summit
10+
link: https://codasip.com/2024/07/16/dac-2024-showcasing-the-future-of-risc-v-through-eda/
11+
---
12+
13+
This week, the 2024 edition of RISC-V Summit Europe took place in lovely Munich, Germany. Those of us who attended last
14+
year’s edition in Barcelona might not have expected the same weather but Munich was up for the challenge and served us a
15+
sunny, hot week, only interrupted by a thunderstorm that shook up some conference attendees (including me!) on the
16+
Wednesday afternoon. However, thunderbolts and lightning were also present in a less literal form as massive applause
17+
following the many talks, and an abundance of photos and selfies taken in the expo hall and at the social events.
18+
19+
![Thumbnail](https://codasip.com/wp-content/uploads/2024/06/Picture1.png)

_layouts/update.html

Lines changed: 10 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -34,6 +34,16 @@ <h2><span class="material-icons">event</span> {{page.date | date: "%d/%m/%Y" }}<
3434
</header>
3535
<article>
3636
{{page.content}}
37+
38+
{% if page.link %}
39+
<a class="button image-button"
40+
href="{{ page.link }}"
41+
rel="noopener"
42+
target="_blank"
43+
title="Continue Reading on Linked Site">
44+
<i class="material-icons">link</i> Continue Reading...
45+
</a>
46+
{% endif %}
3747
</article>
3848
</div>
3949
<div>

0 commit comments

Comments
 (0)