At SC16 there was a discussion between programming language standards for heterogeneous hardware, organised by Khronos. See here for the setup of the session. It was expected to be a heated discussion, but in the end it was a good conversation with lost of learning.

The main message from each language seems to be: “Yes, we’re working on that feature”. This means that a programming language is just like human languages, as new things get named and described world-wide. This also shows the hard work the development of languages bring, as new feature-requests are a constant. Continue reading “Opinions crossing the table: Khronos for world peace” →

To temporarily increase capacity we put Quartus 16.0.2 on an Ubuntu server, which did not go smooth – but at least smoother than upgrading packages to required versions on RedHat/CentOS. While the download says “Linux” and you’re expecting support for multiple Linux breeds, there is only official support for Redhat 6.5 (and CentOS).

Luckily it was very possible to have a stable installation of Quartus on Ubuntu. As information on this subject was squattered around the net and even incomplete, we decided to share our howto in this blogpost. These tips probably also work for other modern Linux-based operating systems like Fedora, Suse, Arch, etc, as most problems are due to new features and more up-to-date libraries than are provided in RedHat/CentOS.

Note1 : we did not install the FPGA on the Ubuntu-machine and neither fully researched potential problems for doing so – installing the FPGA on an Ubuntu machine is at your own risk. Have your board maker follow this tutorial to test their libraries on Ubuntu.

Note 2: we tested on Ubuntu 14.04. No guarantees if it all works on other version. Let us know in the comments if it works on other versions too. Continue reading “Install (Intel) Altera Quartus 16.0.2 OpenCL on Ubuntu 14.04 Linux” →

One of the world’s most used software is far from performance optimised and there is hardly anything we can do about it. I’m talking about Excel.

There are various engine replacements which promise higher speeds, but those have the disadvantage that they’re still not fast enough with really heavy calculations. Another option is to use much faster LibreOffice, but companies prefer ribbons over new software. The last option is to offer performance-optimised modules for the problematic parts. We created a demo a few years ago and revived it recently. Continue reading “Accelerating an Excel Sheet with OpenCL” →

We’re going online with our presentations and tutorials. This makes it easy to reach more people and make our trainings more flexible.

We’re starting with short introductory trainings, but we have bigger plans. Keep an eye on our events (shared on Twitter, LinkedIn, this blog and the newsletter) to see what the offerings are. And you’re very welcome to join!

On 4 October (new date) there will be an OpenCL 101 of two hours for free. Target timezone is East-America and Europe.

Agenda Online OpenCL 101

• Introductions (20 minutes)
  • StreamHPC
  • GPUs and paralellism
  • OpenCL
• By example: Getting started with OpenCL (30 minutes)
• By example: Porting a simple program to OpenCL (30 minutes)
• Q&A in parallel (30 minutes). Ask us any question, for instance:
  • General OpenCL.
  • OpenCL on GPUs.
  • OpenCL on FPGAs.
  • What algorithms work well with GPUs, CPUs and FPGAs.
  • StreamHPC services.
• The next steps (5 minutes).
• Closing words (5 minutes).

Read more here…

Tutorial server

You can already test if the tutorial server works for you by looking around in our demo room. The tutorial itself will be in another room. Use your own name and password “ap“.

The room linked to this resource is not configured correctly.

See you soon!

How water moves through an area given a certain pace of instream, can be fully simulated. We got a request to make such simulation faster, as it took already too much time to do moderate simulations. As the customer wanted to be able to have more details, larger areas and more alternative situations computed, the current performance did not suffice.

The code was already ported to MPI to scale to 8 cores. This code was used as a base for creating our optimised GPU-code. Using a single GPU we managed to get an 44 to 58 times speedup over single core CPU, which is 5 to 7 times faster than MPI on 8 to 32 CPU cores.

For larger experiments we could increase the performance advantage over MPI-code from 7 times to a total of 35 times, using multiple GPUs.

We solved both the weak-scaling problem and the mapping on GPUs

If you add the 9x speedup of the initial performance-optimisation, the total is over 2600x. What could be done in a year, now can be done in 3.5 hours. This clearly shows the importance of software performance engineering. Most code already had some optimisations applied (just like here) and 5 to 7 times speedup is quite achievable.

Read below for some more details. Continue reading “How we sped up a flooding simulation 35 times (from 32-core CPU to multi-GPU)” →

In the past years we have been translating several types of software to AMD, targeting OpenCL (and HSA). The main problem was that manual porting limits the size of the to-be-ported code-base.

Luckily there is a new tool in town. AMD now offers HIP, which converts over 95% of CUDA, such that it works on both AMD and NVIDIA hardware. That 5% is solving ambiguity problems that one gets when CUDA is used on non-NVIDIA GPUs. Once the CUDA-code has been translated successfully, software can run on both NVIDIA and AMD hardware without problems.

The target group of HIP are companies with older clusters, who don’t want to pay the premium prices for NVIDIA’s latest offerings. Replacing a single server with 4 Tesla K20 GPUs of 3.5 TFLOPS by 3 dual-GPU FirePro S9300X2 GPUs of 11 TFLOPS will give a huge performance boost for a competitive price.

The costs of making CUDA work on AMD hardware is easily paid for by the price difference, when upgrading a GPU-cluster.

Continue reading “Get ready for conversions of large-scale CUDA software to AMD hardware” →

The information you find everywhere: on Linux the current “radeon” and “fglrx” are being replaced by AMDGPU (graphics) and ROCm (compute) for HSA-enabled GPUs. As the whole AMD Linux driver team is seemingly working on getting the new and open source drivers ready, fglrx is now deprecated and will not get updates (or very late). I therefore can get to the point:

When using fglrx on Linux, don’t upgrade to Linux distributions with a kernel later than 4.2 or Xorg server versions beyond 1.17!

For Ubuntu this means no 14.04.5 or 16.04 or later. When you have 14.04.4, the kernel will not upgrade when you go to 14.04.5. CentOS/RedHat has such old kernels, there currently is no issue. Fedora users simply have a problem, as they already go towards 4.8.

Continue reading “Dear Linux-users, during the transition period for FGLRX to AMDGPU/ROCm there’s no kernel 4.4 or Xorg 1.18 support” →

On the CUDA page of Wikipedia there is a table with compute capabilities, as shown below. While double checking support for AMD Fijij GPUs (like Radeon Nano and FirePro S9300X2) I got curious how much support is still missing in OpenCL. For the support of Fiji it looks like there is 100% support of all features. For OpenCL 2.0 read on.

Continue reading “CUDA Compute Capability 6.1 Features in OpenCL 2.0” →

From 24 to 28 October we give a 4-day training on OpenCL-on-FPGAs using Altera hardware. The learning goals are correctly writing OpenCL code for FPGAs, learning to work with Quartus and understanding the important optimisation techniques.

The total costs are €2760 excluding VAT for the whole week ( 2 + 2 days of training, one pause day), including a tour in Amsterdam on Wednesday.

See the special event-page for more information.

When we port software to the GPU or FPGA, testability is very important. A part of making the code testable, is getting its functionality fully under control. And you guessed already that run-time generated random numbers takes good attention.

In a selection of past projects random numbers were generated on every run. Statistically the simulations were more correct, but it is impossible to make 100% sure the ported code is functionally correct. This is because there are two variations introduced: one due to the numbers being different and one due to differences in code and hardware.

Even if the combined error-variations are within the given limits, the two code-bases can have unnoticed, different functionality. On top of that, it is hard to have further optimisations under control, as that can lower the precision.

When porting, the stochastic correctness of the simulations is less important. Predictable outcomes should be leading during the port.

Below are some tips we gave to these customers, and I hope they’re useful for you. If you have code to be ported, these preparations make the process quicker and more correct.

If you want to know more about the correctness of RNGs themselves, we discussed earlier this year that generating good random numbers on GPUs is not obvious.

Continue reading “Porting code that uses random numbers” →

In the past year we’ve been working on more internal projects and therefore we’re seeking strong GPU-coders (good OpenCL experience required) worldwide. This way you can combine staying close to your family and working with advanced technologies. You will be on the newly formed international team.

Do understand that we have extra requirements for freelancers:

• You have a personality for working independently.
• You have your own computer with an OpenCL-capable GPU.
• You have good internet (for doing remote access).

We offer a job in a well-known OpenCL-company with various interesting projects. You can improve your OpenCL skills and work with various hardware (modern GPUs, embedded processors, FPGAs and more).

Our hiring-procedure is as follows:

• You send a CV and tell us why you are the perfect candidate.
• After that you are invited for a longer online test. You show your skills on C/C++ and algorithms. You will receive a PDF with useful feedback. (3 hours)
• We send you a GPU assignment. You need to pick out the right optimisations, code it and explain your decisions in detail. (Hopefully under 30 minutes)
• If all goes well, you’ll have a videochat on personal and practical matters. You can also ask us anything, to find out if we fit you. (Around 1 hour)
• If you and the company are a fit, then you’ll go to the technical round. (About 3 hours)
• Made it to here? Expect a job-offer.

We’re looking forward to your application.

Apply for a job as OpenCL expert (freelancer) now!

When you ask how fast code is, then we might not be able to answer that question. It depends on the data and the metric.

In this article I’ll give an overview of different ways to describe speed and what metrics are used. I focus on two types of data-utilisations:

• Transfers. Data-movements through cables, interconnects, etc.
• Processors. Data-processing. with data in and data out.

Both are important to select the right hardware. When we help our customers select the best hardware for their software,an important part of the advice is based on it.

Transfer utilisation: Throughput

How many bytes gets processed per second, minute or hour? Often a metric of GB/s is used, but even MB/day is possible. Alternatively items per second is used, when relative speed is discussed. An alternative word is bandwidth, which described the theoretical maximum instead of the actual bytes being transported.

The typical type of software is a batch-process – think media-processing (audio, video, images), search-jobs and neural networks.

It could be that all answers are computed at the end of the batch-process, or that results are given continuously. The throughput is the same, but the so called latency is very different.

Transfer utilisation: Latency

What is the time between the data-offering and the results? Or what is the reaction time? It is measured in time (often nanoseconds (ns, a billionth of a second), microsecond (μs, a millionth of a second) or milliseconds (ms, a thousandth of a second). When latency gets longer than seconds, its still called latency but more often it’s called “processing time”

This is important in streaming applications – think of applications in broadcasting and networking.

There are three causes for latency:

1. Reaction time: hardware/software noticing there is a job
2. Transport time: it takes time to copy data, especially when we talk GBs
3. Process time: computing the data can

When latency is most important we use FPGAs (see this short presentation on OpenCL-on-FPGAs) or CPUs with embedded GPUs (where the total latency between context-switching from and to the GPU is a lot lower than when discrete GPUs are used).

Processor utilisation: Throughput

Given the current algorithm, how much potential is left on the given hardware?

The algorithm running on the processor possibly is the bottleneck of the system. The metric we use for this balance is “”FLOPS per byte”. This means that the less data is needed per compute operation, the higher the chance that the algorithm is compute-limited. FYI: unless your algorithm is very inefficient, you should be very happy when you’re compute-limited.

The below image shows how the above algorithms on the roofline-model. You see that for many processors you need to have at least 4 FLOPS per byte to hit the frequency-wall, else you’ll hit the bandwidth-wall.

This is why HBM is so important.

Processors utilisation: Latency

How fast can data get in and out of the processor? This sets the minimum latency that can be reached. The metric is the same as for transfers (time), but then on system level.

For FPGAs this latency can be very low (10s of nanoseconds) when data-cables are directly connected to the FPGA-chip. Such FPGAs are on a board with i.e. a network-port and/or a DisplayPort-port.

GPUs depend on how well they’re connected to the CPU. As this is a subject on its own, I’ll discuss in another post.

Determining the theoretical speed of a system

A request “Make this hardware as fast as possible” is a lot easier (and cheaper) to solve than “Make this hardware as fast as possible on hardware X”. This is because there is no one fastest hardware (even though vendors make believe us so), there is only hardware most optimal for a specific algorithm.

When doing code-reviews, we offer free advice on which hardware is best for the target algorithm, for the given budget and required power-envelope. Contact us today to access our knowledge.

Want to help build an important website? OpenCL.org’s components have been designed and partly built, but still a lot of work needs to be done. We’re seeking an intern (or “extern” when not in Amsterdam) who can help us build the site. This internship is not about GPUs!

To complete the tasks, the following is required:

• Technical expertise:
  • HTML5, CSS
  • PHP
  • Javascript
  • jQuery
  • Node.js
  • Mediawiki
  • XSLT
• Can-do mentality
• Able to plan own work
• Good communication-skills
• Available for 3 to 6 months

We don’t expect you know all tools, so we will guide you in learning new tools and techniques. Write us a “email of interest” to info@streamhpc.com, and write what you can and what your objectives for an internship would be.

We’re looking forward to see your letter!

For years we haven been complaining on this blog what AMD was lacking and what needed to be improved. And as you might have concluded from the title of this blogpost, there has been a lot of progress.

AMD is back! It will all come together in the beginning of 2017, but you’ll see a lot of progress already the coming weeks and months.

AMD quietly recognised and solved various totally new problems in HPC, becoming the hidden innovator everybody needed.

This blog is to give an overview of how AMD managed to come back and what it took to get to there. Their market cap supports it, as you can see.

Continue reading “AMD is back!” →

Are you around at ISC and have an opinion on portable open standards? Then you should join the discussion with other professionals at ISC. Some suggestions for discussions:

• (non)-preference for open standards like OpenCL, OpenMP, HSA and OpenACC.
• Portability versus performance.
• Using scripting languages in HPC, like Python.

Below info might change, because a map is never reality! Make sure you check this page or our Twitter channel the evening before.

• Where: ISC in Frankfurt, at the catering area near the lounge areas.
• What: Food, HPC and Open Standards.
• When: Tuesday 21 June, 12:00-13:00

See you there!

Vincent Hindriksen will be walking around at ISC from 20 to 22 June. With me I bring our latest brochure, some examples of great optimisations and some Dutch delicacies. Also we will also have some exciting news with an important partner – stay tuned!

It will be a perfect time to discuss how StreamHPC can help you solve tough compute problems. Below is a regularly updated schedule of my time at ISC.

Get in contact to schedule a meeting.

If you’d like to talk technologies and bits&bytes, we’re trying to make a get-together – date&time TBD.

We have been talking about GPUs, FPGAs and CPUs a lot, but there are more processors that can solve specific problems. This time I’d like you to give a quick introduction to grid-processors.

Grid-processors are different from GPUs. Where a multi-core GPU gets its strength from being able to compute lots of data in parallel (SIMD data-parallellism), a grid-processors is able to have each core do something differently (MIMD, task-based parallelism). You could say that a grid-processor is a multi-core CPU, where the number of cores is at least 16, and the cores are only connected to their neighbours. The difference with full-blown CPUs is that the cores are smaller (like the GPU) and thus use less power. The companies themselves categorise their processors as DSPs or Digital Signal Processors, but most popular DSPs only have 1 to 8 cores.

For the context, there are several types of bus-configurations:

• single bus: like the PCIe-bus in a PC or the iMX6.
• ring bus: like the XeonPhi till Knights Corner, and the Cell processor.
• star bus: a central communication core with the compute-cores around.
• full mesh bus: each core is connected to each core.
• grid bus: all cores are connected to their direct neighbours. Messages hop from core to core.

Each of them have their advantages and disadvantages. Grid-processors get great performance (per Watt) with:

• video encoding
• signal processing
• cryptography
• neural networks

Continue reading “An introduction to Grid-processors: Parallella, Kalray and KnuPath” →

OpenCL header files

When you develop professional software, it is a best practise to have external header files fixed, to have versioning under full control. This way you don’t get the surprises when your colleague has another OpenCL SDK installed. Luckily the Khronos Group has put all version of the OpenCL header files on Github, so you can easily download the targeted OpenCL version.

Download a zip of the header files here:

• OpenCL 1.0
• OpenCL 1.1
• OpenCL 1.2
• OpenCL 2.0
• OpenCL 2.1

If you found problems in one of these, you can directly communicate with the working group by submitting an issue on Github.

OpenCL.lib / libOpenCL.so

But wait, there is more!

You can build your own ICD, as the sources are open (licence). OpenCL version 2.1 is implemented, but it is fully backwards compatible to OpenCL 1.0. You can assume that the vendors use this code for their own, so you can safely use this code in your project.

Get the project from Github.

Want to get an overview of what Heterogeneous Systems Architecture (HSA) does, or want to know what terminology has changed since version 1.0? Read further.

Back in 2012 the goals for HSA were high. The group tried to design a system where CPU and GPU would work together in an efficient way. In the 2013/2014 time-frame you’ll find lots of articles around the web, including on our blog, describing the capabilities of HSA. Unfortunately with the 1.0 specifications most terminologies have been changed.

In March 2015 the HSA Foundation released the final 1.0 specifications. It does not discuss hUMA (Heterogeneous Uniform Memory Access) nor hQ (Heterogeneous Queuing). These two techniques had undergone so many updates, that new terminologies were used.

In this blog post, we’ll present you an updated description of the two most important problems tackled by HSA: memory sharing and task dispatching.

We’ll be tuning the below description, so feedback is always welcome – focus is on clarity, not on completeness.

What is an HSA System?

Where the original HSA goals focused more on SoCs with CPU and GPU cores, now any compute core can be used. The reason was that modern SoCs are much more complex than just a CPU and GPU – integrated DSPs and video-decoder are found on many processors. HSA thus now (officially) supports truly heterogeneous architectures.

The idea is that any heterogeneous processor can be designed by the principles of HSA. This will bring down design costs and enable more exotic configurations from different vendors.

And interesting fact about the HSA-specifications is that it only specifies goals, not how it must be implemented. This makes it possible to implement the specifications in software instead of hardware, making it possible to upgrade older hardware to HSA.

Why is HSA important?

A simple question: “will there be more CPUs with embedded GPU or discrete GPUs?”. A simple answer: “there are already more integrated GPUs than discrete ones”. HSA defines those chips with mixed processors.

CPUs with embedded GPUs used to be not much more than the discrete GPUs with shared memory we know from cheap laptops in the 00’s. When the GPU got integrated, each vendor started to create solutions for inter-processor dispatching (threading extended to heterogeneous computing), course-grained sharing (transferring ownership between processor units) and fine grained sharing (atomics working with all processor units).

The HSA Foundation

Sometimes an industry makes bigger steps by competing and sometimes by collaborating

AMD recognised the need for a standard. As AMD wanted to avoid the problems with introducing 64 bit into X86 and therefore initiated the HSA foundation. The founding members are AMD, ARM, Imagination Technologies, MediaTek, Qualcomm, Samsung and Texas Instruments. NVidia and Intel are awkwardly absent.

Memory Sharing

HSA uses a relaxed memory model, which has full memory coherence (data guaranteed to be the same for all processes on all cores) and is pageable (subsets can be reserved by programs).

The below write-up is heavily simplified to give an overview how memory sharing is designed under HSA. If you want to know more, read chapter 5 from the HSA book.

Inter-processor memory-pointer sharing – Unified Addressing

The most important part is the unified memory model (previously referred to as “hUMA”), which makes programming the memory-interactions in a heterogeneous processor with CPU-cores, GPU-cores and DPS-cores comparable to a multi-core CPU.

Like other modern memory models, HSA defines various segments, including global, shared and private. A difference is that flat addressing is used. This means that each address pointer is unique: you don’t have an address 0 for private and an address 0 for global. Flat addressing simplifies optimisation operations for higher level languages. Ofcourse you still need to be aware that each segment size is limited and there will be consequences when defining larger memory chunks than is available in the segment.

When you have created a memory object and want the DSP or GPU continue to work on it, then you can use the same pointers without any translations.

Inter-processor cache coherency

In HSA-systems global memory is coherent without the need for explicit cache maintenance. This means that local caches are synchronised and/or that caches are shared. For more information, read this blog from ARM.

Fine grained memory – Atomic Operations

HSA allows protecting memory segments to be atomicly accessed. This makes it possible to have multiple threads running on different cores of different processor units, all accessing the same memory in a safe manner.

Small and large consequtive memory segments can be reserved for sharing, from very fine to coarse grained. All threads that have access to that segement are notified when atomic operations are done.

Fine Grained Shared Virtual Memory (HSA compatibility for discrete GPUs)

AMD has done some efforts to extend HSA to discrete GPUs. We’ll see the real advantages with dispatching, but it also works to create a cleaner memory management.

The so called “Fine Grained Shared Virtual memory” makes it possible use HSA with discrete GPUs that have HSA-support. Because it’s virtual and data is continuously transferred between GPU and the HSA-processor, the performance is ofcourse lower than when using real shared memory. You can compare it to NVidia’s Unified Virtual Memory, and it also has been planned to be in OpenCL 2.0 for a long time.

Dispatching

HSA defines in detail how a task gets into the queue of a worker thread. Below is an overview of how queues, threads and tasks are defined and are named under HSA.

Queueing

Before HSA 1.0 we only spoke of “Heterogeneous Queue” (hQ). This is now further developed to “User Mode Queues”. A User Mode Queue holds the list of tasks for that specific (group of) processor cores, resides in the shared memory and is allocated at runtime.

Such task is described in a language called “Architected Queueing Language” (AQL), and is called an “AQL package”.

Agents and Kernel Agents

HSA threads run on one or a group of processor cores. These threads are called “Agents” and come in two variations: normal Agents and Kernel Agents. A Kernel Agents is an Agent that has a User Mode Queue and can execute kernels that work on a segment of memory. A normal Agent doesn’t have a queue and can only execute simple tasks.

If a normal agent cannot run kernels, but can run tasks, then what can it actually do? Here are a few examples:

• Allocate memory, or other tasks only the host can do.
• Send back (intermediate) data to the host – for example progress indication.

If you compare to OpenCL, an agent is the host (which creates the work) and kernel agents are the kernels (which can issue new threads under OpenCL 2.0).

AQL packages: communicating dispatch tasks

There are different types of the AQL (Architected Queueing Language) packets, of which these are the most important:

• Agent dispatch packet: contains jobs for normal agents.
• Kernel dispatch packet: contains jobs for kernel agents.
• Vendor-specific packet: between processors of the same vendor there can be more freedoms.

In most cases, we’ll be talking about kernel dispatch packages.

The Doorbell signal: low latency dispatching

HSA dispatching is extremely fast and power-efficient due to the implementation of a “doorbell”. The doorbell of an agent is signalled when a new tasks is available, making it possible to take immediate action. A problem in OpenCL is the high dispatch times for GPUs without a doorbell – up to the millisecond range, as we have measured. For HSA-enabled GPUs the response-time before a kernel starts running is in the microseconds range.

Context switching

Threads can move from one core to another core – the task will be removed from the current queue and added to another queue. This can even happen when the thread is in running state.

StreamHPC’s position

The solution simply works and makes faster code – we have done a large project with it last year.

It seems that almost the whole embedded processor industry believes in it. AMD (CPU+GPU), ARM (CPU+GPU), Imagination (GPU), Mediatek, Qualcomm (GPU), Samsung and Texas Instruments (DSP) are founders. Companies like Analog Devices, CEVA, Sony, VIA, S3, Marvell and Cadence have later joined the club. Important Linux clubs like Linaro and Canonical are also seen.

The system-on-a-chip only will get more traction, and we see HSA as an enabler. Languages like OpenCL and OpenMP can be compiled down to HSA, so it just takes switching the compiler. HSA-capable software can be written in a more efficient manner, as now can be assumed that memory can efficiently be shared and dispatching new threads is really fast.