Posts by Performation

Do you think spinning rust eats flash's dust? Join the hard drive daddies club

Re: @ChrisBedford - Actually, Hard Disks are reliably unreliable!

Very astute.

Archiving is an issue that is not being addressed.

Currently, digital libraries must survive by copying from one media to another before they deteriorate.

But when you think about it, archiving has always been full of risk. Even paper archives are subject to fire( ref. Alexandria). Even from the earliest days, replication was how knowledge was spread and survived.

So, no matter what media is being used for storage, we will never rely on a single media.

Replication,copying,distribution is how it will have to survive, and we also need to ensure that we still have the methods to read it out while it still is fresh.

Ellison aims his first Oracle 'mainframe' at Big Blue

Re: SPARC T5 - Well Done!

30% better single thread performance than T4 is nice.

However, SUN never ever published any benchmark showing how fast it was on T4.

They announced thatthe T4 was "MUCH BETTER than T3".

However, there was no benchmark published with T3 single thread perf either.

I am tired of these marketing monkeys.

Re: Oracle needs to get these results published when the processor is available...

Well, I have always found TPC-C benchmarks extremely easy to "cheat" on. The IO setup, how thew application is implemented, partitioned, the database tuned etc, plays just as big a role as the capability of the CPUS. So TPC-C is a rather useless benchmark for CPU or even for systems. Too many moving pieces and no way to get apples and apples.

Re: Ibm marketing...

Where is the spec benchmark\ you speak of? I go to spec.org religiously to check for new CPU benchmarks, and Oracle hasn't put any numbers on the board for sparc for years.

Re: Fastest?

To optimize oracle license cost, hunt for the best amd/intel chips with the fewest cores.

I still see very good dual-cores out there. It's too sad Oracle on purpose limits the virtualization options so that there is no viable option on x86. Running Oracle DB on anyrthing but x86 though is . Not only does x86 have the best per core speed, you pay only 0.5 processor license per x86 core.

Re: Fastest?

In my experience coolthread threads give horrible performance. They scale up pretty well within the chip, but each thread is a dwarf snail. So multiplying by the # of threads when those threads are garbage is also not fair.

Re: Fastest?

Well, what does "fast" mean? I've had it with bogus Sun/Oracle marketing.

In my book, the fastest processor is the one with best thread strength.

And Sun/Oracle has had the worst thread strength on the planet for decades.

Cheap Pentium Pro and Pentium II chips used to run circles around sparc in those days. I bet your average cell phone has better thread strength than a coolthreads chip.

Sun was pretty good at scaling up to large systems, and even had the top supercomputer for a short while with the E25k, or was it E10k. But the sun processors have never been impressive. I will pass judgement when I see some industry standard benchmarks. If that ever happens.

Show me the CPU2006 benchmarks

Then I will believe a third of what Sun/Oracle marketing puts out,. As Unix company they were always my favorite, until one day they came to our company with a chief architect and presented future Sun roadmap that showed only coolthreads. Already at this time Sun was lagging the best, but at least they were roughly in the same ballpark as HP/IBM,, and had an edge with the software. I have inherited some of Amdahl's skepticism of the blind reliance of paraleliism to achieve everything. You can use if for most stuff, but not ALL. If I want to go to Europe, I get on a flight with 200 others and it takes 8 hours. With coolthreads, I get on a boat with 1000's of others, and It takes me a week, to get there, but the throughput of the boat is better than that of the flight.

The T5 is a continuation of the T1,T2,T3,T4.

I have always used specint rate benchmarks as a fair and easy benchmark for capacity planning across many different platforms. And in teh few cases where single thread performance matter to ,me, , I have used specint. These benchmark have tracket remarkably well with our application, and seem to be used by most serious industry contenders. Sun (and Oracle ) however, time and time again, have very selectively published benchmarks. While we never would expect specint numbers for coolthread chips to be great, it is very usefui to know. And when Sun an announced T4 with a lot of fanfare, boasting about both throughput and improved thread strength, I was waiting for the CPU2006 numbers. I am still waiting. Sun has published neither specint rate nor specint. My advice to Oracle is to shut up or to publish.