With the biggest data center chip makers banning Russia from next-generation devices, not to mention the exit of handset and software makers from that market, it’s no surprise that Russian researchers are on the fast track to finding ways to circumvent the new technologies that will push the rest of the world.
This is important now in the Russian context, but these efforts will likely spur similar efforts in China, which is also no stranger to tech-flavored sanctions — as we’ve seen in cases like Huawei, for example.
The US government blocked the export of key technologies, including semiconductors, to Russia last week after invading Ukraine. At a minimum, chipmakers complying with US export controls include AMD, Intel, TSMC and GlobalFoundries, with all suspending shipments of product to Russia. Dell, HP and Lenovo have also stopped shipping products to the country, and Oracle and SAP went out of business last night.
The Swiss data center of the financial news service SWIFT has been guarded by the police since Wednesday after important Russian banks were excluded from the system.
The impact of these sanctions cannot be overstated, as the impact is hitting everything from Russia’s most powerful supercomputers to enterprise systems and the much wider world of business and consumer wireless communications.
“Meanwhile,” says Andrei Sukhov, professor and head of the CAD laboratory at HSE University in Moscow, “just naming the problem is not enough; a quick way out of the situation must be sought, relying on the available resources.”
In a recent article on the Association for Computing Machinery (ACM), Sukhov explains how Russian computer science teams are building the next generation of clusters using legacy clustering technologies and a slew of open-source software for managing everything from code portability to parallelization Consider standards like PCIe 3.0, USB 4 and even existing knock-off Russian buses inspired by Infiniband (Angara ES8430).
Such systems would have to be based on what is readily available from a processing and network standpoint, which has more robust native mobile options for Russia (and China, if it matters).
In terms of the highest end of the computer industry, Russia has seven supercomputers in the Top500 ranking, with the 199-node super Chervonenkis being the highest at number 19. Chervonenkis is based on AMD Epyc processors with Nvidia A100 accelerators. There’s not much an integrator can do without key components, including InfiniBand.
As it turns out, the next three most powerful machines (numbers 36, 40 and 43) are similar configurations with AMD processors and Nvidia accelerators.
“Although these devices are not the most powerful, their production is completely independent and does not affect the patent rights of American or European companies,” Suchow explained.
While all the parts could be there, new boards still need to be made, an issue Sukhov says can be circumvented by using wireless protocols as the switching mechanism between processors, although network latency will be below average, which makes it difficult, really Run tightly coupled, low-latency HPC simulations (useful in areas like nuclear weapons simulation, to name just one example).
“Given that the available mobile systems-on-chips are on the order of 100 Gflops, multi-teraflops performance is quite achievable for small clusters of high-performance systems-on-chip,” Sukhov added.
“The use of standard open operating systems such as Linux will greatly facilitate the use of custom applications and allow such systems to be run in the near future. It is possible that such clusters can be heterogeneous, including different systems-on-chip for different tasks (or, for example, FPGAs to create specialized on-the-fly configurable accelerators for specific tasks).”
As he told The Register in a brief exchange following the article: “As far as the existing supercomputers that are already operational, no particular problems are to be expected. These supercomputers are based on Linux and can continue to operate without the support of the companies that provided the hardware and software. According to my information, all scientific supercomputers, even those older than five years, are now operated in normal mode. Only forced control commands or hacker attacks can stop them. But I am not yet aware of such actions in relation to scientific projects, including supercomputers.”
“Of course, it will be impossible to build a new supercomputer in Russia in the coming years. Still, it is entirely possible to meet all current computing and computing needs with the approach we propose. Especially when we apply hardware acceleration to tasks, depending on the type,” he adds.
“It should be noted that our proposed approach is intended for rapid implementation as a pilot project. During this implementation, software solutions and new protocols for data exchange as well as computer technologies are developed.
“In the future, it will be possible to refine the cluster device (for example, to try to launch the release of a new motherboard that will host multiple chips connected by a common bus)” ®