Spectra7 Microsystems Inc V.SEV

Alternate Symbol(s): SPVNF

Technology Semiconductors

Spectra7 Microsystems Inc. is an analog semiconductor company. The Company delivers analog semiconductors at a bandwidth, speed and resolution to enable disruptive industrial design for electronics manufacturers in virtual reality (VR), augmented reality (AR), mixed reality, data centers and other connectivity markets. It creates silicon products that enable copper cables to be longer, thinner, lighter and run at higher performance levels. Its family of products features a patented signal processing technology used in the design of active cables and specialty interconnects in data centers, VR, AR, and other connectivity products. It provides chips, such as HT8181 HDMI 2.0 In-Cable Equalizer, VR8200 Ultra-High-Speed DisplayPort Embedded Interconnect Processor, VR8300 Ultra-High-Speed DisplayPort Embedded Interconnect Processor, VR8050 Interconnect Processor, VR8051 Interconnect Processor, GC2502 Data Center Cable Processor, and GC1122 Dual Channel 112Gb/s PAM-4 Linear Equalizer.

TSXV:SEV - Post by User

Spectra7 Microsystems Inc

Bullboard Posts

Comment by gaugespectreon Apr 02, 2018 4:00pm

82 Views

Post# 27816674

RE:RE:RE:NVDA Servers / data centers

Yes, it will no doubt reduce some of the data center cable demand to have been had if servers were static in development, but I expect a good chunk will still remain. All parts of the cloud are no doubt evolving, incluing sev tech. Others may wish to more precisely interpret some of this forum commentary:

Yojimbo - Friday, March 30, 2018 - link
It depends on the workload. Compute performance is useless if it is heavily constrained by communications bandwidth. The DGX-2 allows every GPU in the node direct memory access at 300 GB/s to all HBM memory in the node (the memory on its own package is still 900 GB/s of course). That's one big pool of 512 GB. Two DGX-1s linked via Infiniband or Ethernet would have two pools of 256 GB (after upgrading to 32 GB V100s). Within one pool (node) each GPU would enjoy only 50 or 100 GB/s of direct memory access bandwidth and between pools the latency for memory operations would be much higher and the bandwidth would be lower than in the DGX-2 where the operations are carried out within one pool.REPLY
mode_13h - Monday, April 02, 2018 - link
You're not wrong. The point of this thing is mainly to efficiently scale up neural network training to larger models than you could fit on the 8 V100's that a DGX-1 can host.

If you don't need more than 8x V100s in a single box, then you probably wouldn't be buying this.REPLY
eSyr - Wednesday, March 28, 2018 - link
DGX-1 has only 8×16 == 128 GB of GPU memory.REPLY
frenchy_2001 - Wednesday, March 28, 2018 - link
Not anymore. It may take some time to trickle to the OEMs, but all V100 are now fitted with 32GB of HBM2 and NV will keep selling DGX1.
So, older DGX1 will hve 16GB/GPU, but newer ones, comparable to the DGX2, will have 32GB.
Biggest difference will be density and fabric. DGX2 allows all 16 GPUs to share their memory.