23 posts • joined 6 Jun 2008
Shoehorning a CPU with 8 memory channels into an existing socket/motherboard infrastructure designed for a CPU with four channels seems like a bit of a challenge!
Maybe be more about cloud servers
The other possibility is that this is about doing something for energy efficient servers in large cloud data centers. I agree with others, don't think ARM is a good fit for XBox unless they plan to do most of the hard work in the GPU and see the processor as more of a conductor than an orchestra.
EMC and x86
Clariion and the most recent iteration of DMX already use x86 processor, Clariion has been based on largely commodity hardware since it's inception.
If you connect the ZinoHD over it's HDMI port, it has no trouble in sending 5.1 or even 7.1 sound to a receiver. I know this for a fact, because I'm doing it.
The reviewer is wrong about the audio
I can only assume that the reviewer is unfamiliar with HDMI connections. If you connect this via HDMI to a TV or better yet a 7.1 receiver it has no problem with digital audio out in 5.1 or 7.1 surround. I know what I'm talking about I just hooked one up.
What about Convey Computer?
I talked to them at SC09 and they had a pretty interesting product that might be a good fit for Intel.
How did Itanium do?
Any news on how Itanium what for Intel's bottom line?
SGI has no containerized data centers?
I'm sure it's just a typo, if not see:
Released version leaks like a sieve
I'm running the released version, and it has a huge memory leak. When it starts, memory usage is around 80MB, after a day of general usage it's closer to 1GB.
I could have sworn Swindon was twinned with LegoLand when I lived there.
Will Power7 fit in existing systems?
Do you know if this will be a forklift upgrade for customers, or will the Power7 be a drop-in replacement in existing systems?
8 threads from 2 cores?
"Joining the 750s will be as-yet-unnumbered 3.2GHz, 3.33GHz and 3.46GHz i5s which contain only two cores, though HyperThreading technology - absent from the first i5 - will present eight cores to the host OS. All three will contain 4MB of L3 cache."
Unless Intel is doing 4 hyperthreads/core, I think that's going to be four cores (or threads) not eight.
Lasers have other advantages vs. Phalanx
The fire control and targetting stuff for a laser is much simpler and probably faster because:
1. The laser beam doesn't drop with distance, so no need to correct for that, which makes getting on target in the vertical axis is simpler
2. It arrives at the target almost instantaneously (i.e. the speed of light vs. a few thousand feet/sec), so the need to account for the target's movement is much reduced, which means that getting on target in the horizontal axis is simpler
All in all, I would suspect that Gatling guns would be rapidly displaced in most applications if this stuff actually works.
I'm not sure about 384 GB
If 384 GB was really needed for virtualization on a two processor system, you have to ask why Cisco didn't use the VMmark benchmark results they published to prove the point. But funnily enough they used the same amount of memory as just about everybody else (96 GB in the form of 12x8GB DIMMs). To me this suggests that memory beyond 96GB for a dual processor is probably a waste unless you have workloads that have a very large memory footprint relative to CPU and I/O requirements.
Nehalem in BX900 Blades
The Dynamic Cube BX900 blade server as announced will never support Nehalem EX, you can't drop Nehalem EX processors into any existing XEON design, it requires a whole new motherboard. It seems unlikely that they would have announced a conventional XEON blade if they expected Nehalem EX to be pulled in by six months as it would be obsolete by time it started to ship. So my guess is that availability date for Nehalem EX is unchanged and that systems will show up in late-Q4 to early Q1.
Match them programmer for programmer...
Makes me thing of Dr. Strangelove, I kpet thinking he was going to mention the mineshaft gap next ;-)
One tester, and suddenly it's trend
I'm part of the W7 private beta (and for every previous beta back to NT 3.1), and I'm very happy so far, does that cancel out chris123. Seemed to me that his post was partly sour grapes that we hadn't received any intermediate builds, he didn't have any specific problems with W7 beta 1 that I could see.
Not much analysis
First, the usual unfounded Observation that an Apple server OS would be more secure than W2K08. Care to justify that a little Tim, it's pretty sweeping statement given that there have been no major security issues with W2K08.
Then there's the whole issue of building an enterprise hardware and software support organization to go along with the servers. Just having the boxes isn't enough, you need a full enterprise support organization or folks will just laugh at you.
Not really aping SVC
The LSI solution is an out-of-band controller and is very limited in terms of additional functionality vs. in-band designs like SVC, DataCore and FalconStor. For example, you'll never see features like thin-provisioning on an out-of-band systems
Need to be specific about what version of Windows!
This might be a more intersting article if it happened to mention what variety of Windows it refers to. For example, if "unpatched Windows" refers to say Windows XP SP1, then this revelation comes in the "duh, no s**t Sherlock" category. If however it refers to say Windows Vista without SP1 installed, then its a tad more alarmainf
Disk vs. memory capacity
While hard drives have always been measured using K=1000 rather than K=1024bits, this will not be true for memory based storage where the building blocks are multiples of 8bits, so 64 x 8 Gbit chips chips is a true 64 GB of storage
Re: Storage systems have caching
Yes, they do, but they are still limited by the number of physical I/Os that the hard disk can handle. Lets say there is an 80% cache hit rate, that means 20% of I/Os have to be satisfied from the hard disks. Now assume 20 drives each capable of 300 IOPs, that 20% can't exceed 6000 IOPs, so overall throughput in real world applications is going to peak at around 30K IOPs, regardless of hw fast the cache is.
If you replace those twenty drives with 10 SSDs each capable of 1000 IOPs and you've got 10,000 IOPs from disk, or 50,000 IOPs overall.
SSD is inevitable, and will ultimately kill what we think of as high performance drives (i.e. 15K RPM FC drives)
- Vid Hubble 'scope scans 200,000 ton CHUNKY CRUMBLE ENIGMA
- Google offers up its own Googlers in cloud channel chumship trawl
- Bugger the jetpack, where's my 21st-century Psion?
- Interview Global Warming IS REAL, argues sceptic mathematician - it just isn't THERMAGEDDON
- Apple to grieving sons: NO, you cannot have access to your dead mum's iPad