Thursday 8 April 2010

New Server Gadgets

iPads? Who needs ‘em, I’m a fan of bigger things. Apologies for the poor quality, I just grabbed a quick iPhone shot of the server’s Task Manager when I noticed just how crazy it looked.

image

See all those little boxes under CPU Usage History – that’s a lot of cores –  48 real cores to be precise on a 4 socket AMD Magny Cours (12 Core Opteron 6100) box that I was using today.

That’s all in a very sleek 2U server that is remarkably quiet. The main memory bandwidth on that is able to move about 20 DVD’s worth of data a sec (that’s a guess but there are 16x1066Mhz DDR3 channels on it so in theory it could push about 130GByte/sec).

The 32\64 Core Intel Beckton (Nehalem EX) based R910 that has just been released looks even crazier as it has 32 Hyperthreaded cores to go with its 32 real cores, and even more insane RAM.

We’re clearly well into the multi-core\multiprocessor era now. While this box isn’t cheap (probably about €15k-20k in this config) it’s still a fairly standard off the shelf product. We probably wont be seeing 48 core desktops in the next year but we will have 12 Core systems by the summer (if they aren’t out there already) and we will have 48 core home systems within a few years. Seems crazy to be able to say that when I’m still using the first multi core PC that I ever bought and it’s not even three years old.

That is provided this whole iPad fad wears off – after all what sort of computer is it if you have no idea how much RAM or CPU it has and how much of that is actually being used? :)

Thursday 1 April 2010

Should we have let the banks just fail..

Last year I wasn’t sure, I certainly thought we should have let Anglo go bust but wasn’t sure about the rest. As someone with almost no assets I’d certainly be no worse off if any of the major banks were to go to the wall but I can see how simply abandoning AIB, BOI and the rest would have been bad for pretty much everyone. However the numbers we’re seeing now really are saying to me that we should have just let them all sink.

The argument that we’re being given is that we had to bail out and recapitalise all the various banks and financial institutions involved in the celtic tiger pyramid scheme because the consequences of letting them fail were too severe. Well were those consequences that we were so afraid of?

For starters we have all the folks who lose a lot of cash they have on deposit. That would be tough alright but that money was deposited with certain guarantees which did _not_ include full protection in the case of the banks being run by incompetant gombeens who couldn’t organise a piss up in a brewery. Interest was paid in return for accepting some (fairly remote) risk, but there was a risk. Still I would feel bad about all of you losing cash and those losses would really grind the economy to a halt. I’d also remind folks that the Government guarantees were not intended to protect your money in these banks so much as reassure the large bond holders that they would be protected. If push came to shove the government would let your money burn, and shrug about it the same way they shrug about the 50% depreciation in the value of your house.

Then we have those institutional investors. Frankly they do this sort of thing for a living and I could care less about them losing money and so should you but he Government seems particularly keen not to let these capitalists face up to some free market realities – you win some and you lose some. They certainly know that banks go bust on occasion and [should]  factor that into their thinking. Those losses would have limited direct impact on the economy but would probably wipe out a fair few pension funds. They would make it very hard for the banks to borrow in future but we’re planning on letting them go bust so that’s no loss. Shame about your pension and it’s quite possible that those potential losses weighed quite heavily in the Governments decisions.

Moving on to the main claim by the Government on why the bailout must happen. In this case there is no sovereign debt default problem (the government isn’t involved at all really) so I can’t see how that should impact Irish Government borrowing but lets assume that markets are irrational and it does mean that international markets treat Ireland Government bonds like those from Greece, or Argentina or any other sovereign in real trouble. At the moment Greek government bonds are trading at 2-3% higher than Irish Government bonds because they are so dodgy. Assume that in this worst case alternative reality we simply retained our pre-existing national debt (€65bn as of  July 2009) and had to pay 4% extra for 10 years to finance all of that. This wouldn’t be the case, of course, most of that debt is already fixed at nice low ~3% rates or such but lets leave it at that to account for debt growth and re-issuance of bonds to cover those that have matures.. Over 10 years the difference in our cost of borrowing as a nation would cost us €43bn more than it would if things had remained static at 2009 levels. That’s assuming we could have actually kept the total debt reasonably under control which seems to be working right now (apart from bank bailout\NAMA funding) even with all our other problems. That’s a pretty pessimistic outcome but it serves as a guideline for “what are we trying to avoid”..

So that’s what we’re trying to avoid. Back in the real world where we have bailed out the banks we’ve kept our borrowing costs under some control but, while we’re not in Greece or Argentina terrain, we have actually taken a 1-1.5% rate hit on the cost of funding our deficits. In addition we’re going to add about €80bn by the end of next year to our extended national debts (forget the off balance sheet\promissory note\deferred blah blah nonsense, someone has to pay to fund these things somehow over the next decade). That comes to around €36bn in interest over our 10 years assuming the extended capital protections for the banks and NAMA disposals break even and are structured in a way that ensures the debt used to cover them does not grow significantly and they are actually eventually paid back in full. Assuming our basic bond rates continue to demand a 1-2% premium then the financing of our pre-existing “normal” national debt will cost us about €6bn more over 10 years. Basically this is costing just as much as the worst case disaster we were trying to avoid and every spare cent we could ever imagine putting together has already been used up on the banks.

Clearly I’ve been quite simplistic here but I think the ballpark numbers are close enough to be well within the margins of error given that all of this depends on fluctuating (and hyptothetical) market rates.

So as I see it right now the bailout is going to cost us almost exactly as much as simply letting the banks fail and having the worst possible things happen as a result would have cost us. We’re still ahead because we’ve protected private and institutional depositors and bond holders within the country but that’s about it. As I see things we’d have been far better off sacrificing Anglo and one or two of the smaller outliers and focusing on protecting core depositors in AID\BOI rather than the protect everything forever approach. Right now I’m certain that the final cost of this bailout will significantly exceed what we would have had to pay to deal with the effects of simply letting Anglo\AIB\IL&P and the rest go under.

Seems like the wrong choice to me.