| AMAX Rackmount GPU servers and systems are fully optimized for the latest GPU computing modules and deliver 10X higher application performance than the latest multi-core CPU systems, providing unprecedented real-world scalability and breakthrough power-efficiency across a wide array of HPC applications. Our wide-range of highly configurable systems' unparalleled performance enable seamless integration of GPU computing for extreme performance, helping scientists and engineers solve the world's most complex computational problems. |
[ POWERFUL, SCALABLE. ]
ServMax™ GPU powered servers are capable of providing up to 11.85 Teraflops of single precision and 3.91 Teraflops of double precision performance – per rack unit! ServMax™ GPU servers offer the perfect GPU density to rack unit ratio to deliver incredible amounts of computational power to crunch numbers faster, solve equations quicker, and predict the weather forecast more accurately. ServMax™ GPU servers are designed to scale as your workload increases and form the building blocks behind our high performance, scalable ClusterMax™ SuperG GPU clusters. That is powerful & scalable. |
 |
|
[ EXPERTLY ENGINEERED. ]
To enable this high level of performance, our team of HPC engineers have carefully taken into consideration the optimal GPU to CPU core ratios, the GPU to RAM ratios, and redundancy if necessary. In addition, system performance is verified via HPL-CTCS (High Performance Linpack) and system-level performance optimizations are made. ServMax™ is expertly engineered. |
 |
|
[ OPEN ARCHITECTURE. ]
All the power and scalability in the world is worth so much more when you’re able to work within an open architecture computing environment. Being open means that we design and customize a solution to meet exact project requirements using the newest, highest performing components the industry has to offer, regardless of vendor. It also means easier maintenance and lower total cost of ownership (TCO). Absolutely nothing is proprietary except AMAX’s dedication to engineering perfection. |
 |
|
|
Xr-4801Gk2 |
|
Xr-2401Gk2
|
|
Xr-2402Gk2 |
 |
|
|
|
|
| # of GPUs |
| 8x Tesla Kepler K20X |
| Theoretical Peak Performance |
11.85 Teraflops (SP)
3.96 Teraflops (DP) |
|
|
| # of GPUs |
| 4x Tesla Kepler K20X |
| Theoretical Peak Performance |
15.80 Teraflops (SP)
5.28 Teraflops (DP) |
|
|
| # of GPUs |
| 4x Tesla Kepler K20X |
| Theoretical Peak Performance |
15.80 Teraflops (SP)
5.28 Teraflops (DP) |
|
|
Xr-1301Gk2 |
|
Xr-1201Gk2 |
|
Xr-4120G |
|
|
|
|
|
| # of GPUs |
| 3x Tesla Kepler K20X |
| Theoretical Peak Performance |
11.85 Teraflops (SP)
3.96 Teraflops (DP) |
|
|
| # of GPUs |
| 2x Tesla Kepler K20X |
| Theoretical Peak Performance |
7.90 Teraflops SP
2.64 Teraflops DP |
|
|
| # of GPUs |
| 12x Tesla Kepler K20X |
| Theoretical Peak Performance |
47.40 Teraflops SP
15.64 Teraflops DP |
|
|
Xr - 2401Gk1 |
|
Xr - 1301Gk1 |
|
|
|
|
|
|
|
| # of GPUs |
| 4x Tesla Kepler K10 |
| Theoretical Peak Performance |
18.32 Teraflops (SP)
.76 Teraflops(DP) |
|
|
| # of GPUs |
| 3x Tesla Kepler K10 |
| Theoretical Peak Performance |
13.74 Teraflops (SP)
.57 Teraflops (DP) |
|
|
|
|
|
 |
