Feeds

Oracle reveals 32-core, 10 BEEELLION-transistor SPARC M7

New chip scales to 1024 cores, 8192 threads 64 TB RAM, at speeds over 3.6GHz

Secure remote control for conventional and virtual desktops

Oracle has revealed details of its next-generation SPARC CPU, the M7.

As John Fowler, Oracle's executive veep of systems predicted when chatting to The Reg last month, the company took the wraps off the M7 at last week's Hot Chips CPU-fest and filled it with goodies to make Oracle software go faster.

Under the hood of the CPU, as depicted in the shot of its die below, are eight “core clusters”. Each cluster contains four cores apiece to reach the 32-core total.

The die of Oracle's SPARC M7 CPU

The die for Oracle's SPARC M7 CPU.
There's TEN BEEELLION transistors on it

There's 64MB of on-chip L3 cache, shared among all the cores and partitioned into 8 x 8 MB chunks. All up, “aggregate bandwidth for the L3 cache on the chip is over 1.6TB/s.”

Oracle says the core clusters share  L2 cache, and that “four cores in the complex share 256 KB of 4 way set associative L2 Instruction Cache, with over 1/2 TB/s of throughput. Two cores share 256 KB of 8 way set associative L2 Data Cache, with over 1/2 TB/s of throughput.”

Inside a Sparc M7 core cluster

Inside a SPARC M7 core cluster

Extensibility is provided by “7 coherence links” per M7, which allows “8 of these chips to be connected together gluelessly.” There's more: Oracle says “32 of these chips can be connected in an SMP configuration. A potential system with 32 chips will have 1024 cores and 8192 threads and 64 TB of RAM.”

Another significant feature is the ability to decompress data at the speed of memory access. Oracle says that will mean the M7 can “directly work on compressed data in memory”. The feature is put into play by a new “M7 Database In-Memory Query Accelerator” that oracle says speeds some database functions.

Oracle is being vague about just when the M7 will land. “2015” is the only guidance on offer at present. But the company is also promising hardware to put the M7 to work in the same year.

Nor is Big Red discussing clock speed, but told former Vulture Timothy Prickett Morgan that the M7 can be expected to clock in faster than the current SPARC M6's 3.6GHz.

At that speed even a single M7 will be capable of mighty deeds.

32 of them in harness might just be dangerously close to Skynet. ®

Internet Security Threat Report 2014

More from The Register

next story
Azure TITSUP caused by INFINITE LOOP
Fat fingered geo-block kept Aussies in the dark
You think the CLOUD's insecure? It's BETTER than UK.GOV's DATA CENTRES
We don't even know where some of them ARE – Maude
Want to STUFF Facebook with blatant ADVERTISING? Fine! But you must PAY
Pony up or push off, Zuck tells social marketeers
Oi, Europe! Tell US feds to GTFO of our servers, say Microsoft and pals
By writing a really angry letter about how it's harming our cloud business, ta
SAVE ME, NASA system builder, from my DEAD WORKSTATION
Anal-retentive hardware nerd in paws-on workstation crisis
Astro-boffins start opening universe simulation data
Got a supercomputer? Want to simulate a universe? Here you go
prev story

Whitepapers

Why cloud backup?
Combining the latest advancements in disk-based backup with secure, integrated, cloud technologies offer organizations fast and assured recovery of their critical enterprise data.
Getting started with customer-focused identity management
Learn why identity is a fundamental requirement to digital growth, and how without it there is no way to identify and engage customers in a meaningful way.
High Performance for All
While HPC is not new, it has traditionally been seen as a specialist area – is it now geared up to meet more mainstream requirements?
Reducing the cost and complexity of web vulnerability management
How using vulnerability assessments to identify exploitable weaknesses and take corrective action can reduce the risk of hackers finding your site and attacking it.
Top 5 reasons to deploy VMware with Tegile
Data demand and the rise of virtualization is challenging IT teams to deliver storage performance, scalability and capacity that can keep up, while maximizing efficiency.