[[ build.model.nick ]]

Choose the price, not the parts. Each model is built with the GPUs, CPU, RAM, and storage that maximizes Deep Learning performance per dollar.

[[ build.model.image_alt ]]

[[ build.model.nick ]]

Basic

OS Ubuntu + Lambda Stack
GPUs 2x NVIDIA RTX 2080 Ti
CPU Intel Core i9-7900X
Memory 64 GB memory
STORAGE 2 TB SSD (OS install)
EXTRA 4 TB HDD

Premium

OS Ubuntu + Lambda Stack
GPUs 2x Titan RTX (Pre-Order)
CPU Intel Core i9-7920X
Memory 128 GB memory
STORAGE 2 TB SSD (OS install)
EXTRA 4 TB HDD

Max

OS Ubuntu + Lambda Stack
GPUs 2x NVIDIA GV100
CPU Intel Core i9-7920X
Memory 128 GB memory
STORAGE 4 TB SSD (OS install)
EXTRA 4 TB HDD

Customize

Not seeing what you want?

Add a Protection Plan

3-year protection for [[ getWarrantyPrice(true) ]]

[[ getSubtotal(true)]]
Talk to an engineer
(650) 479-5530

About the [[ build.model.title ]] Basic

GPUs

GPUs are the most critical piece of hardware for Deep Learning. The Basic has 2x RTX 2080 Ti GPUs (Turing Architecture). For Deep Learning in 2018, the 2080 Ti offers the best price/performance trade-off of any GPU on the market. Each 2080 Ti has 11 GB of memory and 13.4 TFLOPs of FP32 performance (the standard precision for Deep Learning training). For most Machine Learning tasks, 2x 2080 Ti will perform about 80% as fast as 2x Quadro GV100 and Tesla V100. We estimate that 2x 2080 Ti will perform about 95% as fast as the Premium's 2x Titan RTX.

Processor

During training, the CPU preprocesses data and feeds it to the GPUs. A slow processor will cause the GPUs to waste cycles waiting for this data. Core count and PCI-e lane count are important CPU performance factors. More cores means faster data preprocessing; more PCI-e lanes means faster transmission of that data to the GPUs. The Premium uses Intel's Core i9-7900x (10 cores, 44x PCI-e lanes). Its core-to-GPU ratio is 5, which follows the best practice of at least 1 CPU core per GPU. The Premium's CPU, combined with its PLX-enabled motherboard, provides 16x PCI-e lanes to each GPU (the max possible).

Motherboard

A motherboard's PCIe lanes have a significant effect on Deep Learning training performance. PCIe lanes are data pipes connecting the GPUs and CPU. The number of PCIe lanes connecting each GPU to the CPU varies from 4x to 16x, depending on the motherboard. More PCIe lanes is better because more lanes = more throughput. The Basic's motherboard has PLX chips, which ensures that each GPU gets 16x PCIe lanes (the maximum possible as of 2018).

Memory

A Deep Learning computer should have at least as much RAM as GPU memory. According to this rule, a machine like the Basic, which has 2x NVIDIA 2080 Ti GPUs, should have at least 22 GB of memory (2 GPUs * 11 GB memory per GPU). The Basic has 64 GB of memory, which provides plenty of overhead. If you work with large data sets (e.g. many large images), consider upgrading to Premium, which has 128 GB of system memory.

Storage

Most datasets do not fit in RAM. In such cases, during model training, subsets must be repeatedly swapped in and out of RAM from nonvolatile storage. Such a pipeline requires fast, solid state storage; without it, the GPUs would waste cycles waiting for their next batch of data. The Basic was designed with this constraint in mind; it has two nonvolatile storage devices: a 2 TB solid state drive (fast) for data you're training on now, and a 4 TB hard disk drive (slower) for everything else. Files located in the /data directory are stored on the HDD; all other files are stored on the SSD.

Network

Network interface speed is irrelevant to Deep Learning performance, unless you're doing multi-node distributed training (in which case, you'll need at least 40 Gbps), or pulling training data from network attached storage. With 1 Gbps ethernet, this desktop's network interface is far faster than most ISPs provide. If your training data is stored on a remote server on the same LAN, consider upgrading to Max, which has 10 Gbps network speed.

Who bought a Basic?

About the [[ build.model.title ]] Premium

GPUs

GPUs are the most critical piece of hardware for Deep Learning. The Max has 2x Titan RTX GPUs (Turing Architecture). Each Titan RTX has 24 GB of memory. We estimate that the Titan RTX will be about 5% faster than the 2080 Ti and 85% as fast as the NVIDIA GV100 and Tesla V100. The Titan RTX is well suited for training with large batch sizes and large networks.

Processor

During training, the CPU preprocesses data and feeds it to the GPUs. A slow processor will cause the GPUs to waste cycles waiting for this data. Core count and PCIe lane count are important CPU performance factors. More cores means faster data preprocessing; more PCIe lanes means faster transmission of that data to the GPUs. The Premium uses Intel's Core i9-7920x (12 cores, 44x PCIe lanes). Its core-to-GPU ratio is 3, which follows the best practice of at least 6 CPU core per GPU. The Premium's CPU, combined with its PLX-enabled motherboard, provides 16x PCIe lanes to each GPU (the max possible).

Motherboard

A motherboard's PCIe lanes have a significant effect on Deep Learning training performance. PCIe lanes are data pipes connecting the GPUs and CPU. The number of PCIe lanes connecting each GPU to the CPU varies from 4x to 16x, depending on the motherboard. More PCIe lanes is better because more lanes = more throughput. The Premium's motherboard has PLX chips, which ensures that each GPU gets 16x PCIe lanes (the maximum possible as of 2018).

Memory

A Deep Learning computer should have at least as much RAM as GPU memory. According to this rule, a machine like the Premium, which has 2x Titan RTX GPUs, should have at least 48 GB of memory (2 GPUs * 24 GB memory per GPU). The Premium has 128 GB of memory, which provides plenty of overhead. The 128 GB of memory available on the Premium is quite standard for researchers who work with large data sets (e.g. many large images).

Storage

Most datasets do not fit in RAM. In such cases, during model training, subsets must be repeatedly swapped in and out of RAM from nonvolatile storage. Such a pipeline requires fast, solid state storage; without it, the GPUs would waste cycles waiting for their next batch of data. The Premium was designed with this constraint in mind; it has two nonvolatile storage devices: a 2 TB solid state drive (fast) for data you're training on now, and a 4 TB hard disk drive(slower) for everything else. Files located in the /data directory are stored on the HDD; all other files are stored on the SSD.

Network

Network interface speed is irrelevant to Deep Learning performance, unless you're doing multi-node distributed training (in which case, you'll need at least 40 Gbps), or pulling training data from network attached storage. With 1 Gbps ethernet, this desktop's network interface is far faster than most ISPs provide. If your training data is stored on a remote server on the same LAN, consider upgrading to Max, which has 10 Gbps network speed.

Who bought a Premium?

About the [[ build.model.title ]] Max

GPUs

GPUs are the most critical piece of hardware for Deep Learning. The Max has 2x NVIDIA GV100 GPUs (Volta Architecture). Each GV100 has 32 GB of memory. Our Deep Learning benchmarks show that the GV100 is comparable in speed to the Tesla V100; it is approximately 40% faster than the 2080 TI for Deep Learning training tasks. We estimate that it will be 5-10% faster than the Titan RTX for Machine learning training tasks.

Processor

During training, the CPU preprocesses data and feeds it to the GPUs. A slow processor will cause the GPUs to waste cycles waiting for this data. Core count and PCIe lane count are important CPU performance factors. More cores means faster data preprocessing; more PCIe lanes means faster transmission of that data to the GPUs. The Premium uses Intel's Core i9-7920x (12 cores, 44x PCIe lanes). Its core-to-GPU ratio is 6, which follows the best practice of at least 1 CPU core per GPU. The Premium's CPU, combined with its PLX-enabled motherboard, provides 16x PCIe lanes to each GPU (the max possible).

Motherboard

A motherboard's PCIe lanes have a significant effect on Deep Learning training performance. PCIe lanes are data pipes connecting the GPUs and CPU. The number of PCIe lanes connecting each GPU to the CPU varies from 4x to 16x, depending on the motherboard. More PCIe lanes is better because more lanes = more throughput. The Max's motherboard has PLX chips, which ensures that each GPU gets 16x PCIe lanes (the maximum possible as of 2018).

Memory

A Deep Learning computer should have at least as much RAM as GPU memory. According to this rule, a machine like the Max, which has 2x NVIDIA GV100 GPUs, should have at least 64 GB of memory (2 GPUs * 332 GB memory per GPU). The Max has 128 GB of memory, which provides plenty of overhead. The 128 GB of memory available on the Premium is quite standard for researchers who work with large data sets (e.g. many large images).

Storage

Most datasets do not fit in RAM. In such cases, during model training, subsets must be repeatedly swapped in and out of RAM from nonvolatile storage. Such a pipeline requires fast, solid state storage; without it, the GPUs would waste cycles waiting for their next batch of data. The Max was designed with this constraint in mind; it has two nonvolatile storage devices: a massive 4 TB solid state drive (fast) for data you're training on now, and a 4 TB hard disk drive (slower) for everything else. Files located in the /data directory are stored on the HDD; all other files are stored on the SSD.

Network

Network interface speed is irrelevant to Deep Learning performance, unless you're doing multi-node distributed training (in which case, you'll need at least 40 Gbps), or pulling training data from network attached storage. With 10 Gbps ethernet, this desktop's network interface is far faster than most ISPs provide. If your training data is stored on a remote server on the same LAN, the Max's 10 Gbps network interfaces is the proper choice.

Who bought a Max?

[[ component.name ]]

[[ option.description ]] [[ build.getPriceDiff(component, option)]]