Puget Systems News
One of the questions I get asked frequently is "how much difference does PCIe X16 vs PCIe X8 really make?" Well, I got some testing done using 4 Titan V GPU's in a machine that will do 4 X16 cards. I ran several jobs with TensorFlow with the GPU's at both X16 and X8. Read on to see how it went.
Posted in Featured Systems on 05/21/2018
Our Machine Learning / AI Workstation configurations are single CPU, multi-GPU and optimized for deep-learning workloads. These are great platforms for working with Tensorflow, Caffe and Torch as well as other frameworks like CNTK, MXNet, etc.. We have two base recommendations to meet performance and budget requirements.
Receive the best hardware for your workload without a large upfront investment! Puget Systems has partnered with LiftForward to redefine ownership with business leasing, providing a 24-month subscription option, or device-as-a-service (DaaS).
"I don't know where to begin." That's what I told Beth, who was working the counter at my local auto parts store. I sheepishly placed a sleek can refrigerant on the counter and began to explain my predicament. I purchased a refill because I already owned a pressure gauge, I assumed it was compatible with any can. I found out that isn't the case when I tightened the gauge and heard the stem pop.
PhotoScan makes heavy use of both the central processors (CPUs) in a computer and the video cards (GPUs) to run many of the calculations involved in turning still images into a 3D model or map. Intel's new Xeon Scalable processors offer configurations with dozens of CPU cores, as well as the ability to support multiple GPUs - so let's see how they perform in PhotoScan.
Intel has launched their new Xeon Scalable processor series, with very high core counts and multi-CPU configurations. How do they stack up to single-socket workstations using other Intel and AMD processors when rendering in V-Ray?
"Mac or PC?" - the age-old question among computer enthusiasts. How fast are Apple and PC workstations when rendering in V-Ray? And which offers a better value?
"Mac or PC?" - the age-old question among computer enthusiasts. How do Apple and PC workstations compare for content creation and rendering in Cinema 4D?
Intel has launched their new Xeon Scalable processor series, with very high core counts when used in dual CPU configurations. How do they stack up to single-socket workstations using other Intel and AMD processors in Cinema 4D?
After Effects users are often held back by the performance of their workstation, yet a surprising number of users lock themselves into the Mac ecosystem. In this article we will be looking at just how much faster a PC workstation can be in After Effects compared to the iMac Pro and Mac Pro.
Apple may have had a stranglehold on video editing workstations for many years, but with 4K, 6K, and even 8K footage being used more and more, many are starting to jump ship in favor of a PC workstation. Most people know that they can get more out of a PC, but just how much faster is a PC versus a Mac Pro or iMac Pro in Premiere Pro?
With the rise of 4K, 6K, and even 8K footage, colorists using DaVinci Resolve are quickly discovering that the Mac ecosystem is simply not able to give them the performance they need. Most people know that they can get more out of a PC, but just how much faster is a PC versus a Mac Pro or iMac Pro?
Want to see how your system stacks up to the latest hardware? Download our free, public Pix4D benchmark tool which will walk you through a couple of basic projects and display the calculation times. We've also included a video walkthrough, showing how to use this tool, and some comparison results from recent workstations built here at Puget Systems.
PhotoScan makes heavy use of the central processor (CPU) in a computer to run many of the calculations involved in turning still images into a 3D model or map. Different steps in that process utilize the CPU in various ways, though, so we are looking at how several Intel and AMD processors compare in this application.
In this article, we discuss an issue we have been seeing with the Asus X99E-10G WS motherboard and triple or quad NVIDIA GPU configurations. When running a GPU and CPU heavy workload, it will cause the system to crash and give the DPC_WATCHDOG_VIOLATION error or "blue screen of death". We have included an instructional video outlining a work around we have found to resolve this issue. We are also currently coordinating with NVIDIA on a permanent fix. Updates will be published to the article as they become available.
I attended the Microsoft Build 2018 developers conference this week and really enjoyed it. I wanted to share my "big picture" feelings about it and some of the things that stood out to me. I'm not going to give you a "reporters" view or repeat press-release items. This is just my personal impression of the conference.
I have been qualifying a 4 GPU workstation for Machine Learning and HPC use. The last confirmation testing I wanted to do was running it with TensorFlow benchmarks on 4 NVIDIA Titan V GPU's. I have that systems up and running and the multi-GPU scaling looks very good.
PhotoScan makes use of both the CPU and GPUs (video cards) in a computer, during different steps of the photogrammetry workflow. One of the configuration options within this program also allows the CPU to be utilized during steps that are primarily performed on the GPU - and it is enabled by default. However, we have found in our testing that this option usually hampers performance more than it helps!
PhotoScan makes use of the video cards in a computer to assist with the computation of certain steps. As such, both the model of video card used and the number of GPUs present in a system can have an impact on the amount of time those steps take. In this article, we take a look at how multiple GeForce GTX 1080 Ti cards scale in performance across a few different CPU platforms.
We have recently found an issue with Premiere Pro 2018 version 12.1.X with ProRes footage causing up to 35% longer export times and a 50% drop in live playback performance with both ProRes 422 HQ and ProRes 4444 footage. Unfortunately, currently the only fix appears to be rolling back to version 12.0.1 of Premiere Pro 2018.
Microsoft generously wrote in the ability to install a FEW Linux distros but unfortunately, it doesn't work right out of the box. In the guide I will discuss how to install the WSL and get the distros installed.
GPU Memory Size and Deep Learning Performance (batch size) 12GB vs 32GB -- 1080Ti vs Titan V vs GV100
Batch size is an important hyper-parameter for Deep Learning model training. When using GPU accelerated frameworks for your models the amount of memory available on the GPU is a limiting factor. In this post I look at the effect of setting the batch size for a few CNN's running with TensorFlow on 1080Ti and Titan V with 12GB memory, and GV100 with 32GB memory.
PhotoScan makes use of the video cards in a computer to assist with the computation of certain steps. As such, the model of video card used can have an impact on the amount of time those steps take. In this article, we take a look at the GeForce 1000-series - based on NVIDIA's Pascal GPU architecture - to see how they compare to each other.