Calxeda’s first ARM server is a serious threat to x86 server domination
Fifteen months ago, a CPU developer named Calxeda made waves when it announced a joint effort with HP to develop dense ARM servers that would challenge x86 supremacy in the server market. The company promised that it could leverage the low power consumption of ARM products to build clusters of Cortex-A9 SoCs inside a rackmounted chassis.
Anandtech’s Johan De Gelas (a name old-timers will recognize from Aces Hardware) has benchmarked and written the first review of a Calxeda-based system, the Boston Viridis. This system contains six EnergyCards, totaling 24 CPUs (96 Cortex-A9 cores) clocked at 1.4GHz. Anandtech ran the system through a range of synthetic and real-world application tests and compared its single- and quad-threaded performance to both Atom and Xeon-based solutions.
The results are sure to make Intel sit up and take notice. The ECX-1000 processor at the heart of the Viridis lags even Atom in some metrics, like bandwidth utilization (Atom is ridiculously slow compared to Xeon processors, just to put that in perspective). Its per-thread performance in integer workloads, however, is quite competitive with Intel’s in-order architecture. While it never matches the Xeon-based products in terms of single-threaded performance per clock, the synthetic tests show the ECX-1000 is an excellent product.
The real-world tests are stunning. Not only does Calxeda’s array of “wimpy” cores outperform Xeon processors in web server tests, it beats them in both raw performance and performance-per-watt. De Gelas writes that “the Calxeda’s ECX-1000 server node is revolutionary technology.” After seeing the performance figures, I agree. There’s a place for ARM products in the datacenter. This also makes AMD’s long-term bet on an ARM server solution look like a good idea.
What's your take on HPQ Project Moonshot as a Caldexa ally/competitor? From what I've read, HPQ kicked off the project with Redstone based on Caldexa-designed ARM chips, but since then has moved on to Gemini (announced for Q2/13 launch) that reportedly uses INTC Atom chip. Interestingly, TXN may also be jumping on board the HPQ program with its own design ARM chip.
Will HPQ, DELL, or Quanta, with their huge financial and R&D resources, steal the dense microserver market away from Caldexa and AMD-Seamicro with designs based on INTC chips? Even if ARM chips proved to be a competitor to INTC in the space, I'd expect a large company like TXN to develop ARM chips superior to those of Caldexa and AMD in short time.
The word on the street is that HP will use the TI KeyStone II chips in its second-generation "Gemini" Moonshot servers, which the company previewed last summer. During that preview last June, the only processor that HP talked about was Intel's "Centerton" Atom S1200 server chip, which was announced in December last year, and it never mentioned ARM processors, not even once. (Funny that.)
HP has not provided much in the way of feeds or speeds for the Gemini machines except that they will use the two-core Atom S1200 processor, which has 64-bit processing, supports VT virtualization assist, and ECC scrubbing on main memory and is certified to run server variants of Linux and Windows.
It has not said, either, how the Gemini machines will stack up compared to the "Redstone" Moonshot boxes that HP launched in November 2011 using the 32-bit Calxeda ECX-1000 ARM chips, which include an on-chip distributed Layer 2 switch that is very clever.
Uneducated BS. Who in their right mind compares a 24 socket server with a 2 socket server which is less than half the price ?! I know Johan (and Charlie Demerjian) from the Aceshardware days and respect his work but he slipped up here. The correct comparison would have been against a 4 socket Romley server which is what it was designed for or against greater nodes of 2-way LV Xeons. A 4 socket Romley would have easily beat this Boston 24 node A9 server on performance, performance/watt and with the right hardware configuration, price.
Looking at the big picture, there now exists competition for Intel in servers. That's bound to drive down ASPs and ARM based startups like Boston don't need much to survive in a race to the bottom. As startups, they don't have market capitalization to worry about like the old established Intel.