Re: Ahhhh. The Guardian
If you really believe that Oh dear.............................
138 posts • joined 25 Oct 2012
If you really believe that Oh dear.............................
Well said Boltar.
analysts are like rating agencies, useless
No blame accepted might equal $10Million, but not $100Million.
HP = Muppets
Anyone can create a server business, so if makes sense for EMC to do so, (similar to Cisco), they just go to a Chinese ODM, or Supermicro and viola server bus' without HP dummies being involved.
I'm not sure the cultures are the biggest problem, after all a big chunk of HP cam from it's merger with Digital, Waltham MASS.
Arrogance is HPs biggest problem.
Para 3, yawn
2 hears ago I think?
So you know ALL in the city? Silly statement
Groan, yep only Tory MPs tell fibs!
Emperor Meg you are without clothes.
Nope not the only game in town.
50,000 cores. 256TB RAM, NumaScale
Falconidae is the family name rather than a "term"
And the first Supercomputers
Nope normal Wifi is not the way to go, but there is an unbreakable Wifi in development
1.75Gbps half-duplex 802.11ad
The inflection point will not be towards a hybrid flash array such as Nimble, but straight to standard server based SDS, such as Nutanix.
The hybrid boys will get squeezed from incumbents, (EMC, Netapp, et al), at the top and Nutanix, Symplivity, etc from the bottom.
On train Wifi wont work for long. You need APs along the lines
Correct, correct and Correct.
That's what we have developed
Data packages are expensive, free Wifi isn't!
As Wifi is connection orientated, with long latency delays, it drops connection frequently. Put some tunnels and cuttings in the way and packets get buffered for so long then overflow and loss of service.
The solution is a connection orientated Wifi where the APs are beside the track, with 100usec latency between APs, result unbreakable Wifi at 100mph.
Looking for funding, serious!
I thought iWARP was long dead in the water?
Isn't Thunderbolt an I/O interface?
The Intel announcement is an interconnect for inter-processor distributed memory applications, like CFD.
HPC boxes are clusters of individual dual or quad socket Linux boxes each running its own OS.
SGI Altix UV is not a cluster, (distributed memory), but is a ccNUMA shared memory platform. ScaleMP the same, although without HW aided cache coherency lookups.
Old Sun E-series, were true SMPs, i.e. all CPUs connected by a crossbar switch.
Almost all modern day multi-socket servers are ccNUMA platforms.
Why in earth would it be as significant as that?
The 360 catapulted IBM from ANO mainframe maker to #1 by a country mile. Think, transistor, 360, semiconductor, CMOS, network. This is just an evolution of the mainframe
Metropolitan Wifi everywhere solves the problems.
Its because for PWRs Areva is the only game in town. Thorium much cheaper in comparison, but try explaining it to a politician!
I thought Hitachi had built a 10Megawatt prototype Flouride salt reactor, but could be wrong?
"This year the rain started in March and stopped last week. "
where do you live? Obviously not the UK.
Bleeding edge those Fuji boys eh?
We do NOT store excess sugars as fat, we store them as Glycogen, a sugar polymer, like starch or lignin, (wood).
All fooods need enzymes to be digested easily, not just fats, amilases, (carbs), lipases,(fats), proteases.
The highest energy content of foods is 1. fat, 2, Protein, 3. Carbs.
Eating high a protein diet "should" be good as the excess amino acids are excreted through the kidneys and the excess sugar mols are stored as glycogen.
Hadvar, I like you, but no polar bears in the antarctic
Good stuff closer 60%?
FFS that's dangerous levels of OH, not for drinking me old mucker, not for long anyway
Thorium Flouride reactors were developed in the fifties, but didn't go mainstream as they don't make Plutonium.
As regards not being ideal ,that would because dumbo politicians need a quick fix, aka Areva, (except they can't build and commission a PWR in less than 20 years.
Thorium salt reactors would be a great tech for the UK to specialise in, IMHO :)
It's also bloody expensive and dog slow if you go the PWR Areva route, (1970s technology). If Sizewell is built in 10 years from now, then I'll eat my hat. Now Thorium Flouride, that appears a sensible plan
And the shameless Pachauri is an accountant I believe, who used to work for the Indian railways.
Erm, because you have to put up with Sparc?
Herring gulls must be at least twice the size of Black headed gulls, you sure youve got this right?
Lesser black backed gulls similar to Herring Gulls
Replacing ALL tier 1 spinning stuff with flash is a non-sense, as you can't get around the sustained write problem that all flash has versus DRAM.
Hmm VMAX out Violin in?
So I put my enterprise b******s on the line for a 9 year old startup, who are still burning cash and who have no OEMs, no thanks.
Only 2 years out aren't they?
Thorium reactors should be in production now for Hinckley Point instead of a 40 year old PWR design form Areva.
It's NOT a Tokamak, that's the point
No the first use is as a district wide energy plant, i.e small - medium sized town.
The spin on jet engines is at the moment you have to use an oil based fuel and the current order books for both Boeing and Airbus are gangbusters, so looks interesting to Mr Wallstreet
No the benefit is its also cheap. NB cheap is not free, but compared to Fission reactors it should be virtually free.
Its big but doesn't work for long and still massively energy negative. Mind you it employs more people in France.
Maybe the problem is the Tokamak as stated by LM
The previous, (1-2 years ago), video from LM said they specifically went far away from the Tokamak and instead for eliptical sphere design, as it was much easier to keep the plasma away from the vessel wall
Sounds like the beginnings of a Bob Dylan song
General Atomics fail
A spokesman for General Atomics said it was the they had heard of this development?
Then he or she is a lazy dummy, as I read and watched a video about this in 2011 or 12
I am skeptical they keep asking for investment mind you!!
At 1500 cores and 8TB RAM, it's not even on the same planet as the MetOffice IBM supers, but it has 8TB of global memory.
Mind you you can get more cores and much more RAM for a fraction of the cost using NumaScale ccNUMA HBAs.
NumaConnect can connect together 1000s of servers into a single system image, with up to 256TBytes of RAM and its cheapish!!!