Gordon Haff Quotes (49 Quotes)

Over time, the performance-per-watt equation is becoming increasingly important. Its power is relative to performance as opposed to raw power or lowest power consumption.

Sun has given a speed boost to its installed base as the company focuses on x64 systems and the forthcoming Niagara processor,

It's not like Dell had been making investments in Itanium and suddenly decided it wasn't going to do that and pull back its support. It had a relatively older product it wasn't promoting at all, and it really doesn't have a near-term path where it could move forward if it wanted to.

If developers don't make the transition to Apple and Intel, they'll probably be making the transition to Windows and Intel. Intel doesn't have a lot to lose.

Certainly it's at least at the same level of sophistication that Office 97 was in terms of features and capabilities, which is enough for most people, ... Microsoft seems determined to see how much money it can squeeze out of buyers, but in doing so, they're opening up an opportunity for somebody else to come in.

Those do not appear to be cherry-picked numbers of some particular benchmark on a blue moon with a tailwind blowing.

I see this as fine-tuning more than changing direction. Obviously, they're struggling financially, but in general this is a small number of people in the scope of things at Sun.

People have tried a lot of special purpose processing devices over the years and, with the exceptions of graphics units and arguably floating point units, general purpose processors have always won out in the end.

Microsoft has built its business on working with third party software vendors and wants to leverage that expertise to the supercomputing arena.

In the sense of having this ongoing awareness of and feedback mechanism, it does mirror certain elements of open source development. If not openness of code and standards, there is greater visibility into what's going on. Customers have visibility more than an ability to necessarily influence Microsoft greatly.

AMD has made some gains in mobile technology, but has been only marginally successful. There is a higher level of integration with notebooks than with desktop PCs, which gives Intel an edge since they are focused on a complete mobile platform.

Intel has inconsistently put energy into storage. It hasn't had a real comprehensive strategy in the past.

Obviously, Virtual Iron is a startup in a space where a lot of people are playing. I'm not about to predict that they're going to take the market by storm. On the other hand, they have some unique capabilities that they can build into a broader market.

Certainly Power Everywhere, to some level, includes OEMing systems, though for the most part it's more focused on the use of the Power architecture by various partners in client devices, game consoles, whatever, than it is in computer systems, ... Power Everywhere and OpenPower are really about using collaborators to get Power into new types of devices.

A fast chip got faster. This certainly helps to cement their strong current position.

This is a bigger story for Apple than for Intel. Apple liked Intel's roadmap better than that of Power PC, and the idea is that they will now sell a higher percentage of notebook computers.

You have to do a lot of running just to stay in place in this business.

Software patents are a mess, ... Even if you don't believe in abolishing them entirely, there have been far too many examples of patents being granted even in light of overwhelming evidence of prior art and of dusty old patents on their third owner suddenly being used for a form of blackmail.

You don't want to see things slip, but at the end of the day, if Montecito gets delivered per this revised schedule and the performance increase is in the range of what Intel has talked about a two-times increase, it will still give OEMs like HP a competitive high-end processor offering.

The problem is, to be at that nirvana, Apple would have to have the ISVs increase their investment on Apple hardware, at least in the transition. That is very difficult for Apple.

Basically, no one ever used hardware-based IA-32 execution, so better to use the silicon for something else. Of course, basically no one uses software-based emulation either, but at least that doesn't cost chip real estate.

I don't think Microsoft had much choice in the matter. The bottom line is there is not a whole lot of market to buy that product from Microsoft.

Business desktops are a narrow, narrow margin business they are almost a commodity. You have to be very efficient in order to make money selling desktops.

People do not typically leave companies if they are deliriously happy with their roles, but sometimes even people who are reasonably content with their jobs find better opportunities that come along.

In 2004 and 2005 we saw new technologies and approaches enter the server market as spending and growth began to resume. In 2006, we shouldn't expect to see so many brand-new things, but rather the expansion and maturation of products and technologies already in the market.

Sun has to keep up with the other guys. IBM is going to be coming out with the Power5 soon. The reality is in this industry you have to put a lot of energy into running to stay in place. What happens if you don't make these kinds of advances is you fall behind. So this upgrade is in some respects about maintaining parity with the competition.

Programming Cell is relatively hard. Certainly a higher-level, more abstracted model makes programming a lot easier.

Intel woke up from its frequency delusion and is now going full-bore on dual-core.

doesn't fundamentally change UltraSparc and Sun's positioning in any way. It's the type of incremental upgrade that computer makers sort of have to deliver.

The new chipset should help them maintain a leadership position in mobile processor technology.

The potential here is a lot more efficient use of hardware, better centralized management, and potentially more flexibility with users in that they're not tied to a particular desktop device. I don't believe there's a one-size-fits-all approach, we'll probably end up with a combination of thin clients and various types of fat clients.

By the middle of next year, we are going to be in a situation where Intel will have erased some of the functional gaps that they have had with AMD, ... By erasing the gap, Intel certainly has the opportunity to bring their much larger and better resourced manufacturing and development organization to bear.

Certainly, the fact that you need to be using brand new hardware is a selling disadvantage out the door.

Intel has shown preliminary benchmarks that backup their general statements about performance. Based on what I have seen, there is not reason to think the products will not live up to their billing.

It's probably not possible to be head and shoulders above the competition in the x86 space. There just isn't that much opportunity for being compellingly different. That said, they're nice boxes.

It's still not clear how visible that will be. I don't think it will be pervasive in 2006.

Microsoft is going to be very late to the game and will need to overcome a presumption that it will favor Windows.

The fact that Itanium didn't take off as expected is really because it is only in certain types of applications such as databases that there was a driving need to go beyond 32-bit to the larger memory you see in RISC servers today. But 64-bit extensions to x86 are here, and there is no penalty to having a server that is capable of supporting 64-bit, whether you need it or not. That's the nice thing about the x64 approach.

This isn't a new generation of servers aimed at grabbing a lot of new-to-Sun business. That's what the Galaxy server it just announced is for. But this is a good upgrade.

Intel let a competitor sneak into a market where they weren't before. Still, it was hard to believe that AMD would be able to maintain the kind of performance lead that they have enjoyed over the past one to two years.

A number of companies are starting to look at supercomputing for various data analysis applications.

Microsoft has made a few unsuccessful attempts to become a key player in the supercomputer market but the company could be in a better position for success this time.

Enterprises must either adopt multi core or stick with slower chips.

I assume Scott will paint it as things are back on track and the company is in good hands, but I think the reality is he would have preferred to have waited until the financial results were really there.

It's a local company with some useful code, engineers who obviously have skills and background and experience in the type of systems management that HP has. We don't know the terms of the deal, but it probably was not all that expensive, ... It looks to be a good deal for HP.

History says that binary translation basically doesn't work. The day may come when someone can do a good enough job with it, but that concept has been thrown out there many times in the computer industry, and it's always fallen flat on its face.

I don't think anybody knows exactly what the end result of open sourcing the chip technology is.

With the benefit of 2020 hindsight, there would not be an Itanium today.

Obviously, they have to sell lots of systems of this size to pay back what has to be a considerably larger RD investment than they have even with Galaxy.