AI Reinvented – NVIDIA GTC 2018

Last week, NVIDIA kicked off its annual GPU Technology Conference with a keynote speech from Jensen Huang, founder and CEO of NVIDIA, where he unveiled a series of advancements NVIDIA has made towards the future of real-time rendering and artificial intelligence. These announcements included the new high-end professional-grade Quadro GV100 GPU, a doubling of the memory for the Tesla V100 GPU accelerator, and the new DGX-2 Machine Learning supercomputer with scalable architecture based on NVIDIA’s NVSwitch low-latency networking fabric.

The new Quadro GV100, designed to be world’s most powerful workstation GPU, is the first Quadro graphic card to feature the Volta architecture, with NVIDIA RTX ray tracing technology. Equipped with 640 Tensor Cores, the GV100 delivers extreme memory capacity, scalability and performance for Deep Learning, rendering, VR, and AI applications. Quadro GV100 is now available integrated inside AMAX’s powerful Deep Learning workstations, the 2x GPU DL-E200 and 4x GPU DL-E400, to deliver the best-in-class DevBox for studios, start-ups and labs.

The upgraded Tesla V100 delivers 2x boost in local memory capacity from 16GB to 32GB, with the goal to tackle the most memory intensive Deep Learning and HPC applications. The additional memory now enables data scientists to overcome CPU-to-GPU memory transfer bottlenecks, and train complex Deep Learning models with a 50% increase in performance and efficiency. The new Tesla V100 32GB is now available with all of AMAX’s Award-Winning Deep Learning servers. Each system is configured and optimized to meet each customer’s compute requirements, stage of development and performance goals.

The most shocking announcement at GTC 2018 was NVIDIA’s DGX-2, the most powerful supercomputer so far. Weighing 350 pounds, the DGX-2 is world’s first 2-PFLOPS Deep Learning system, packing 16 of the latest SXM3 V100s (featuring 32GB memory) to accelerate the newest AI models that were previously impossible to train. According to NVIDIA, a DGX-2 system can train FAIRSeq, an advanced language translation model developed by Facebook, in less than two days – a 10x improvement in performance compared to DGX-1. The DGX-2 is now open for preorders through AMAX, and is targeted to ship in Q3 on a first order/first ship basis, so get your spot now! On-site training and installation with be provided by AMAX as an official NVIDIA Elite Partner.

For GTC this year, AMAX showcased its dedication in providing the best-in-class Deep Learning platforms to fast track AI development at any scale—development to hyperscale production. AMAX offers the most comprehensive line of GPU-integrated solutions optimized for Deep Learning—from high-performance workstations for development, high-density GPU servers, to large-scale production-ready GPU clusters, delivering unmatched performance-per-dollar and performance-per-watt level. Being a one-stop solution provider for all Deep Learning development needs, AMAX’s highly technical and experienced engineers work closely with our customers to analyze specific requirements and performance goals to recommend the optimal platform and configuration to fit their needs. To learn more about AMAX’s systems featured at GTC 2018, please visit Or schedule your technical consultation today by emailing and putting “GPU Solutions Technical Consult” in the subject line.

Posted in AMAX News, Big Data / Hadoop, Cloud Computing, Data Center, Deep Learning, Engineering, Enterprise Computing, GPU Computing, HPC Computing, Internet of Things, Tradeshow/Events | Tagged , , , , , , , , | Comments Off on AI Reinvented – NVIDIA GTC 2018

Six Reasons to Step-Up to DGX-1 with Volta – Because AI Doesn’t Wait

NVIDIA® DGX-1™, the most powerful Deep Learning computing engine on the market, is inspired by the demands of AI and data science to fast track Deep Learning development out-of-the-box. Among the worlds’ first portfolio of AI supercomputers, the DGX-1 accelerates Deep Learning workflow from desktops, to data centers and into the cloud. For those of you who are currently using the DGX-1 with Tesla P100 GPUs, based on the Pascal architecture, here are six reasons why you should upgrade to the AI “Fast Lane” with Volta.

Reason 1: Record-Breaking Performance

DGX-1 Volta delivers 3x the performance of DGX-1 Pascal by packing 8x Tesla V100 GPUs to provide 1 petaFLOPS of “data-center-like” compute power in one enclosure, using only 3500W power and only 3 RU of space. Currently being the world’s most advanced data center GPU to accelerate AI, the Tesla V100 GPU delivers 125 Tensor TFLOPS by leveraging 5,120 CUDA cores and 640 Tensor cores. With its performance-engineered Deep Learning software stack, DGX-1 delivers 30% faster training than other GPU-based systems.

Reason 2: The Power of Tensor Cores

While many solutions offer GPU-accelerated performance, DGX-1 unlocks the full potential of Tesla® V100 GPUs, NVLink™ and the new Tensor Core architecture. With 640 Tensor cores, Tesla V100 is the world’s first GPU to break the 100 TFLOPS barrier of Deep Learning performance. The next generation of NVIDIA® NVLink™ connects multiple V100 GPUs at up to 300 GB/s, for 10x the I/O performance of PCIe connected GPUs, to create the world’s most powerful computing servers. AI models that would consume weeks of computing resources can now be trained in a few days.

Reason 3: Training and Inference in One Platform

Google has stated that 7ms is an optimal latency target for real-time inference-based workloads, and Tesla V100 GPUs are able to deliver 3x more throughput performance when compared with P100s, and 40x more when compared to a CPU-only server within the 7-millisecond latency budget.

Reason 4: Your DGX-1 is Faster and Smarter with Volta

The DGX Software Stack taps into the best of Volta technology. Engineers can now benefit from faster training cycles using Volta optimizations for leading Deep Learning frameworks, including Caffe, Caffe2, MXNet, PyTorch and TensorFlow, Theano and Torch. The NVIDIA Collective Communications Library (NCCL) 2.1 Volta-optimized version delivers over 90% multi-node, multi-GPU scaling efficiency for Deep Learning training.

Reason 5: Accelerated ROI on Deep Learning

Accelerated training means faster time-to-insights and lower operation expenses on non-value add effort. A recent study by Forrester Research found that companies using DGX-1 saw an average of $2.4M in incremental revenue, and $436k saved in data scientist productivity.

Reason 6: Volta Upgrade Promotion – Take Action Now!

Get onto the AI “Fast Lane” at an unbelievable price. Now through March 30th, 2018, purchase the DGX-1 Volta Upgrade kit for only $39,200 (20% discount.) As an Elite member of the NVIDIA Partner Network Program, AMAX is ready to deliver on-site training and installation, in addition to NVIDIA DGX System Support. With your DGX-1 Volta upgrade kit, no change to existing power supply will be needed. All installation can be completed in as fast as 3 hours, with system performance and operation validated.

All this sound amazing but you don’t even have a DGX-1? Contact us immediately to set up a consultation on how the DGX-1 can fast track your AI development and accelerate your ROI. For more information about the DGX portfolio of AI supercomputers or about how to sign up for the Volta upgrade, email us at The full catalog of AMAX’s award-winning Deep Learning solutions can be found at

Posted in AMAX News, AMAX Services, Data Center, Deep Learning, Enterprise Computing, GPU Computing, HPC Computing, Promotion | Tagged , , , , , | Comments Off on Six Reasons to Step-Up to DGX-1 with Volta – Because AI Doesn’t Wait

AMAX Releases [SMART]Rack P47 Deep Learning & Rendering Solution at SC17

Last week at SuperComputing Conference 2017, AMAX released the [SMART]Rack P47, a PetaFLOP-In-A-Rack solution featuring AMD CPUs and GPUs, designed for deep learning and rendering applications.

The [SMART]Rack P47 combines the power of 20x AMD EPYC™ Processors and 80x AMD Radeon Instinct™ GPUs, based on AMD’s revolutionary “Vega” architecture, to deliver one PetaFLOP of 32-bit FP single precision compute power at a stunning 30 GigaFLOPs/Watt. It is designed to address today’s most complex computational problems in Artificial Intelligence, Deep Learning, VDI, advanced rendering, compute and research at an unmatched performance-per-dollar and performance-per-watt level.


The [SMART]Rack P47 PetaFLOP-Performance-in-a-Rack solution made its first appearance at SIGGRAPH 2017 in July, when AMD CEO Lisa Su proudly unveiled the rack and described it as “the most beautiful server rack that has ever been built.”

The supercomputing-class performance of the fully-integrated rack is achieved by 20x ServMax™ P47 servers, designed with high scalability to support a variety of workloads such as virtual desktop infrastructure, deep learning, machine learning, advanced rendering and compute. Each ServMax™ P47 server supports 1x AMD EPYC™ CPU and 4x AMD Radeon Instinct™ MI25 PCIe GPUs, delivering 20,480 Stream Processors, 49.20 TFLOPS single-precision, and 98.40 TFLOPS half-precision performance. Integrated with AMD’s ROCm open software platform and MIOpen libraries, the ServMax™ P47 servers deliver superior performance with unmatched manageability for deep learning training deployments in datacenters.


The [SMART]Rack P47 solution also integrates advanced memory and storage technologies; unified 100Gb InfiniBand/ Ethernet HPC fabric for increased in-rack bandwidth and productivity; integrated rack level out-of-band management layer for remote server health monitoring and rack orchestration; and a high-density rack cooling system add-on for optimal efficiency in heat exchange.

“As a high-performance technology provider enabling enterprises to close the gap between scale up performance, compute density and cost, AMAX sees the [SMART]Rack P47 as a game changer,” said Julia Shih, VP of Business Development, AMAX. “Starting from a single ServMax™ P47 server, we can scale upwards to supercomputing-class performance by leveraging AMD EPYC CPUs, AMD Radeon Instinct, and AMD ROCm development tools to support Deep Learning, rendering, and a host of other applications.”

AMAX is now taking pre-orders for both the ServMax™ P47 server and the fully-integrated [SMART]Rack P47, with delivery estimated in Q1 of 2018.

Posted in AMAX News, Cloud Computing, Data Center, Deep Learning, Enterprise Computing, GPU Computing, HPC Computing | Tagged , , , , , | Comments Off on AMAX Releases [SMART]Rack P47 Deep Learning & Rendering Solution at SC17

Fortifying AMAX’s Commitment to Industry-Low Failure Rates: How Silicon Valley’s Leading OEM Appliance Manufacturer Achieves Quality

One of the distinguishing factors that keeps customers returning to AMAX as an OEM Appliance Manufacturing partner is how well we put our products to the test.  Seriously, we mean business!  Guaranteeing that products never leave our facility without being put through a battery of tests is our way of protecting your brand and bottom line. Quality translates into a rock solid, reliable product, ensuring fewer warranty claims, less downtime and disruption to your end users, and an overall better customer experience of your brand. For nearly 40 years, AMAX has built a reputation of upholding the highest standards, giving ISVs, corporations, and data centers peace of mind knowing every system AMAX ships will be of the highest quality.


How can we speak so confidently?  Years and years of perfecting our manufacturing and testing process.

Executing Quality

  1. AMAX Testing Automation (ATA) – ATA is AMAX’s proprietary test suite designed to flag significantly more component errors than standard tests. Does this mean our parts are faultier? No. It means our standards are exponentially higher than industry standards, and anything less than perfect we consider a failure and will not ship out. ATA can be customized into a very specific test program using applicable diagnostics to evaluate performance against precise function requirements. It is able to capture all test and device logs for simplified data collection. Most uniquely, ATA is capable of testing large volumes simultaneously, allowing us to achieve aggressive manufacturing schedules while still maintaining quality.
  2. AMAX ISO 9001 5-Stage Quality Gate – our thorough 5-Stage quality process begins with testing individual components prior to manufacturing, followed by assembly based on centralized manufacturing process instruction to ensure accuracy and consistency of builds, then our proprietary ATA high-temperature burn-in stress test, Functionality Validation Testing (FVT), and a final QA inspection.
  3. Optional Add-On Tests – because our test programs are entirely customizable based on our customers’ needs, we have additional tests that can be administered during production: Highly Accelerated Life Test (HALT), Design Verification Test (DVT), and Ongoing Reliability Test (ORT).

img_1553bBut it doesn’t stop there.  Recently, AMAX has invested in several Environmental Stress Screening (ESS) chambers to stress systems through a range of temperature settings. Industry reliability numbers have shown that simply doing a room temperature test or only a slightly elevated heated test will not catch all faulty or potentially weak components.  Therefore, a combination of both hot and cold stress tests in a controlled environment is the only way to ensure that all integrated components are thoroughly tested and likely to perform as needed in the field. With a temperature range between -30°C to over 100°C and at a 10°C/minute ramp rate, the new ESS chambers allow us to better simulate a variety of testing situations. A few more specifics about the ESS chambers:

  • Testing at cold temperatures will find any marginal connectivity issues in the system
  • Warmer temperatures will identify any network failures for borderline high speed signal devices
  • The ESS chambers are averaging about an 85% yield with most devices passing; however they are catching an additional 15% of failures that were not captured with earlier testing
  • Capacity of the ESS chamber is 40x 2U units at full capacity and up to 70x 1U units, depending on the rack design

ESS testing is an add-on service for customers who must have the most reliable and optimally-performing systems in the field. Once ESS testing is complete, systems progress through final stages of the 5-Stage Quality Gate into Final Test, where proper customer images and settings are verified, and logs are deleted before packing.

As many of today’s leading technology companies rely on AMAX to build the server and rackscale appliances that feature their company name, any quality or hardware performance issues can throw a wrench in customer adoption or negatively affect their brand. This is why AMAX strives to deliver platforms that are reliable beyond industry standards, and offers a slew of value-add services (Custom Branding, New Product Introduction Program (NPI), Global Logistics, etc.) to help its partners succeed fast and grow quickly.



Interested in partnering with us for your OEM appliance needs?  Learn more about our OEM Appliance Program here. We can’t wait to work with you!

Posted in AMAX News, AMAX Services, Engineering, ISV Appliances, Server Appliance Manufacturing, Server OEM, Total Computing Solutions | Tagged , , , | Comments Off on Fortifying AMAX’s Commitment to Industry-Low Failure Rates: How Silicon Valley’s Leading OEM Appliance Manufacturer Achieves Quality

Introducing Our Newest AMAX Family Member, Heart-Melting Specialist

On Monday, June 26th, a member of the AMAX family arrived to work and heard a weak, high-pitched cry coming from beneath a car. Upon investigation, he found a straggly gray ball of fur, which turned out to be a malnourished and terrified 4-week old kitten. A severe eye infection had left him blinded with both eyes sealed shut.


They say it takes a village to raise a child, and it was no different with this little guy. The AMAX family immediately jumped into action, warming him with a towel, cleaning him up, and feeding him kitten formula from a tiny baby bottle.


After a visit to the vet deemed him healthy outside of the eye infection, the AMAX team has spent the next weeks nursing him back to health.

We are proud to introduce the newest member of the AMAX family: Matrix GPU On-Premise Cloud (Powered by Bitfusion Flex), or Neo for short.


We hope we can keep him as the AMAX office cat and mascot, as his presence has melted the hearts of all who have nursed him and held him sleeping in their laps, and brought a new level of camaraderie, compassion, and joy to the office.


If you would like to donate to the care of little Neo, we hope you will consider procuring one of our MATRIX Deep Learning Solutions. Not only is MATRIX the best solution on the market for AI development and deployment, but proceeds go to keeping a roof over our little guy.collage1b

Posted in AMAX News, Deep Learning, GPU Computing, Product Development | Tagged , , , | Comments Off on Introducing Our Newest AMAX Family Member, Heart-Melting Specialist