4x GPUs = 3.5x faster training in TensorFlow, Keras, PyTorch, Caffe, and Caffe 2. The Lambda Quad is designed specifically for multi-GPU training. GPU-GPU communication throughput is significantly increased through a PLX-switched PCI-e topology. Now available with AMD Ryzen Threadripper.
Customizable to 4x GTX 1080 Ti, Titan Xp, or Titan V
Intel Xeon E5-1650 v4 (6 cores)
Upgradeable to Intel Core i9-7920X (12 cores)
Upgradeable to 128 GB
Upgradeable to 4 TB
Upgradeable to 8 TB RAID 5 (3x 4 TB)
250x faster training than a normal computer
Deep Learning requires specialized hardware. Normal CPU-based computers are far too slow. We benchmarked the Lambda Quad against a number of high-end CPU-based workstations on a representative sample of training tasks. For example, when training ResNet-50 on ImageNet, the Lambda Quad is...
faster than an Intel Core i9-7900X desktop
faster than an Intel Xeon W-2133 desktop
faster than an Intel Xeon E5-1650 desktop
faster than an Intel Core i7-7820X desktop
Pre-Installed with Every Framework
Easy Upgrades with Lambda Stack
Lambda Stack is a software tool for managing installations of TensorFlow, Keras, PyTorch, Caffe, Caffe 2, Theano, CUDA, and cuDNN. If a new version of any framework is released, Lambda Stack can manage the upgrade, including updating dependencies like CUDA and cuDNN. Learn more.
We're experts in Machine Learning
When you chat with us, there's no sales talk. We're a small team of experts who take pride in our machines. Call us, tell us what you're working on, and we'll talk you through your options.
QUAD is our four GPU workstation custom tailored for deep learning research and engineering teams. Four NVIDIA GeForce GPUs lets your team iterate and deliver results four times faster than a single GPU.