Is Samsung preparing for a push into the server space? That’s the (probably) entirely justified conclusion of PC World’s Agam Shah, who noted that Samsung’s decision to license the new Cortex cores from ARM in October likely means that Samsung is considering a run at the server market, either with its own hardware or via ARM-based cores for third-party hardware. Naturally, Samsung isn’t commenting. “Samsung is a lead partner of ARM's new Cortex A50 processors. However, we're not in a position to comment on our plans for how we'll use the Cortex A50 as part of our Exynos product family," Lisa Warren-Plungy, a Samsung Semiconductor spokeswoman, wrote in an e-mail to Shah. ARM has announced both the Cortex A57, designed for high-end, single-thread application processing, and the A53, which the company claims to be the most power-efficient 64-bit processor—not to mention the physically smallest it’s ever produced. Customers can license both designs. Although the chips will run ARM’s 64-bit v8 architecture, they will also function as 64-bit chips. AMD announced it would incorporate the design into its own ARM-based Opteron chips for the data center. Broadcom, Calxeda, HiSilicon, Samsung and STMicroelectronics were the other five licensees. Although Samsung does not make server hardware, it might not need to; the company’s semiconductor manufacturing prowess is considerable, and the company recently invested in building a new chip line for 20nm and 14nm processes, spending $1.9 billion in the facility. That would put Samsung in the same league as Intel, which has gained considerable market share on the strength of its manufacturing alone. Of course, the ARM-based microserver market remains unproven, and Samsung would likely have to develop a server-class SOC, not just manufacture the core itself. Still, developing low-power, high-performance ARM silicon for tablets and phones is something Samsung understands well. In related Samsung data-center news, the company announced at the end of last month that it had begun producing two enterprise-class solid-state-disc (SSD) drives for the data center. Available in capacities of 120 gigabytes (GB), 240GB and 480GB, the SM843 drive is a multi-level chip (MLC) SSD offering that can offer 11,000 write IOPS (inputs outputs per second), with up to 70,000 (sustained) random read IOPS. The related SM1625 uses a dual-port SAS interface, features up to 23,000 (sustained) random write IOPS, and randomly reads up to 101,000 IOPS when using both ports. The SM1625 ships in densities of 100GB, 200GB, 400GB and 800GB. This week, Intel announced the latest generation of its own SSD family, the S3700. Image: wavebreakmedia/Shutterstock.com Is Samsung Eyeing the Data Center?
Is Samsung preparing for a push into the server space? That’s the (probably) entirely justified conclusion of PC World’s Agam Shah, who noted that Samsung’s decision to license the new Cortex cores from ARM in October likely means that Samsung is considering a run at the server market, either with its own hardware or via ARM-based cores for third-party hardware. Naturally, Samsung isn’t commenting. “Samsung is a lead partner of ARM's new Cortex A50 processors. However, we're not in a position to comment on our plans for how we'll use the Cortex A50 as part of our Exynos product family," Lisa Warren-Plungy, a Samsung Semiconductor spokeswoman, wrote in an e-mail to Shah. ARM has announced both the Cortex A57, designed for high-end, single-thread application processing, and the A53, which the company claims to be the most power-efficient 64-bit processor—not to mention the physically smallest it’s ever produced. Customers can license both designs. Although the chips will run ARM’s 64-bit v8 architecture, they will also function as 64-bit chips. AMD announced it would incorporate the design into its own ARM-based Opteron chips for the data center. Broadcom, Calxeda, HiSilicon, Samsung and STMicroelectronics were the other five licensees. Although Samsung does not make server hardware, it might not need to; the company’s semiconductor manufacturing prowess is considerable, and the company recently invested in building a new chip line for 20nm and 14nm processes, spending $1.9 billion in the facility. That would put Samsung in the same league as Intel, which has gained considerable market share on the strength of its manufacturing alone. Of course, the ARM-based microserver market remains unproven, and Samsung would likely have to develop a server-class SOC, not just manufacture the core itself. Still, developing low-power, high-performance ARM silicon for tablets and phones is something Samsung understands well. In related Samsung data-center news, the company announced at the end of last month that it had begun producing two enterprise-class solid-state-disc (SSD) drives for the data center. Available in capacities of 120 gigabytes (GB), 240GB and 480GB, the SM843 drive is a multi-level chip (MLC) SSD offering that can offer 11,000 write IOPS (inputs outputs per second), with up to 70,000 (sustained) random read IOPS. The related SM1625 uses a dual-port SAS interface, features up to 23,000 (sustained) random write IOPS, and randomly reads up to 101,000 IOPS when using both ports. The SM1625 ships in densities of 100GB, 200GB, 400GB and 800GB. This week, Intel announced the latest generation of its own SSD family, the S3700. Image: wavebreakmedia/Shutterstock.com