Smaller, cheaper, less profitable machines -- what is Apple thinking?! They should totally try to be Dell.
Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Apple to Use Newly Released Quad-Core Desktop Processors?
- Thread starter MacRumors
- Start date
- Sort by reaction score
Smaller, cheaper, less profitable machines -- what is Apple thinking?! They should totally try to be Dell.
Why did noone reply to the above post?
If the xmac/midi/iMac pro is not going to happen, and surely it isn't, then were still in the world of compromise (no i7).
The Q9000 is $348, similar to the $369 for the Q9550S. Sure the Q9550S is faster (2833 MHz vs 2000 MHz, bigger cache, faster FSB), but 45W TDP seems more like the compromise style of the all-in-one.
Not a bad idea....
Seriously, nobody is suggesting that Apple should produce a dozen models to cover every range. They'd only need one or two, or three max to fill that hole.
A premium mini-tower and a small-form-factor using desktop parts, starting at $899 for a quad Core 2 with integrated graphics.
Smaller, cheaper, less profitable can be turned to be:
Versatile, value-priced, and selling in much higher volume.
Sorry, but not everyone needs a 2000+ dollar workstation for general use. They just don't.
I certainly don't.
Some see the need for All-in-one. Good for them. Not everyone does. Some want big screens, some want two screens, some want NO screens.
And not everyone wants the compromises with laptops, when the machine isn't going to go anywhere.
There is a huge hole, bigger than ever.
An xMac, Mac Midi, or whatever they want to call a headless small factor desktop machine is not going to impinge on those who want MacPro.
It probably won't impinge much on those who want a nice AIO form factor and small desktop footprint.
What it will do, is help people buy Apple products rather than buying Dell, HP, or other, even home-built machines. Even if the profits aren't as hugely high, isn't it better than turning good customers away, without an option?
And it is a bogus argument that it is too niche. What is Apple if it isn't a hugely successful company at putting niche technology into the mainstream.
Secondly, a modest/moderate desktop machine is most company's bread and butter, not their niche machines. Most hardware companies are not going broke by offering mainstream desktop hardware, and it would not amount to Apple spreading itself too thin. If Apple is spread too thin, it isn't by computer variation.
And, BTW, the car company analogy doesn't work, either. Their problem isn't variation, it is that the products don't match the demand, or the expectation of quality for the price. And their costs are out of control, and their decision making is entrenched in bureaucracy, so improvement is difficult to come by.
Apple is also not meeting a demand, although their quality is pretty good, with a few niggles in terms of details. Their decision making is starting to be a little cloudy, if their detail niggles are telling about the decisions going on.
What rumored possibility of 30" iMacs? There's no evidence that Apple has any 30" backlit-LED screens on the horizon.
Then they found out its much easier to make more efficient multi-core designs than it was it focus only on clocks speed. There's a hard limit to most designs in the 3.5-4ghz range.
Uh, yes can. Especially, with the Nehalem designs, Bloomsfield and Gainestown are the exact same core. The only difference is the second QPI link for communicating with the second CPU. If the Core i7 Bloomsfield is going to suck, then so is the Xeon Gainstown.
Sad but true. Apple seems to think everyone exists in one of the two extremes. If you're a consumer, you have very minimal needs. If you're a professional, you have extremely high needs. If you live in reality where the line no where near that neat and tights, you're out of luck. You either have to deal with hardware that is either does not meet your needs or is too expensive or you have to try to not be driven insane by windows on the machine you really want.
Not much of a market for such a machine unless it also acts as a TV. 30" displays are too large and expensive for consumers and all in one designs aren't the best fit for professionals.
Invariably, it is different in different countries.
What doesn't help matters is that trying to use Dell's website to find comparable hardware is a nightmare.
As such, there's lots of opportunites to make a mistake, or to be dishonest and try to game the system to force it to the conclusion that you want.
This all merely makes it harder to get to the real question that is the underlying objective of these sorts of inquiries: for reasonably equal configurations where we've made the honest effort to zero out the obvious differences, and from a value paradigm also tried to avoid known price traps (eg, overpriced Apple RAM), what's the answer to the question of relative product value?
And there are always going to be limitations in 'equal' comparisons. For example, in post #143, orhun offers the dual-quad Xeon Dell Precision T7400. This isn't a perfect match, but one that's hopefully close enough: if you look at the specs, you'll see that its using the 667MHz RAM (thus, 1333MHz FSB) whereas the Mac Pro uses 800Mhz (1600MHz FSB), which is a ~20% difference in bandwidth, which probably shouldn't be ignored...but has to at least be noted as a mismatch.
I'm not suggesting that this is a poster child example, but I've found that there's a lot of extremely bad hardware combinations out there, which mismatch between the CPU and the Motherboard. The crux of the problem are vendors cutting costs. They make what can almost be called a "Bait & Switch" by dropping a new (often multi-core) CPU onto an older, cheaper, CPU plug-compatible Motherboard. They hope is that the buyer only pays attention to the CPU and their price, and overlooks that the old motherboard lacks the bandwidth/pipes to let the new CPU run without being bottlenecked. As such, you end up with a system that has great looking specs, but doesn't ever deliver the expected real world performance.
Personally, I did the exercise of a full, honest, and objective cross-comparison back in April 2008. It came up again last month and I did a quick spot check and found that nothing significant had appeared to have changed. There is the new i7 CPU, but I reject any use of that hardware as a basis of comparison: this is strictly a Xeon vs Xeon comparison by explicit objective and intent.
Agree that the progress towards leveraging multiple cores has appeared slow. For the most part, it has been in some specialty software that has been explictly designed for it that the benefits exist. This is essentially why we need to avoid claims like "one Chip i7 is equivalent to two...", since benchmarks are merely guidelines and not an assurance that a particular software solution is actually optimized or not. Besides, benchmarks have also been known to have been 'gamed' by some hardware manufacturers, so they can't be 100% trusted.
-hh
I checked the SPECrate numbers (which measure multi-core performance) and the quad 2.66 GHz Dell Core i7 outperforms the octo-core Mac Pro 2.8 GHz on multi-core jobs !!
At one third the price.
Guess that really shoots a hole in the price/performance equation.
Not at all, unless your job is to run SPECrate numbers all day
I'm not trying to downplay the i7 CPU as being "bad" or anything. I'm merely pointing out that it is different and we don't necessarily know what all of the implications of that difference really mean.
Similarly, I've been around long enough to have seen hardware vendors purposefully "game" their products to post better benchmark numbers. I'm not saying that it is happening here, but merely pointing out that it has happened before. In any event, benchmarks are merely generalized predictors of expected real-world performance, for when integer and floating point operations are actually distilled down into the actions that the computer's operator will see as better responsiveness (or whatever) that result in the operator being able to get more done...higher productivity.
Plus, this sort of shifting is merely an expected part of technology advancements. Give Intel a fwe more months and the core technologies of the i7 will be in their Xeon counterpart. Maybe prices will come down too, but who knows? That's expected noise and flux from living in a non-static R&D environment.
-hh
PS: another factor in price-performance is the Law of Diminishing Returns. It always costs a disproportionate amount more to get that last 5% of performance. Yes, this often can mean a literal doubling in price...and sometimes more. If you want the biggest bang for the buck, you're never going to buy at the top, but instead, you're going to look for the "knee in the curve". Suffice to say that the Mac Pro (...and all of the other dual-quad Xeon's being sold...) are above the knee. This does say something positive in regards to the xMac advocacy, but its problem remains that its a slice-of-a-slice-of-a-slice marketshare.
In general, I prefer the fewer hassles that stem from initially having an OEM Warranty, so I tend to follow the "buy new and keep it forever" paradigm.
Insofar as specifics, I don't care for buying off of eBay either...as far as I'm concerned, its mostly become a "Seller's Market", not a Buyer's Market. This isn't specific to computers, but pretty much anything on eBay.
For Apple, I don't have any particular qualms about buying refurbs, although the price-value determination has to be there for it, since I generally expect to be keeping the machine for a long time, so saving $500 when it is going to be amortized across 5+ years isn't as big of a deal. My current desktop Mac is 5.5 years old and counting; the one that I had before that went 7+ years which was IMO "too long", but I was waiting for the G5 CPU.
-hh
now that these chips are out...
...how long until they're in new Apple machines? Does anyone know the historical delay? a week? a month?
...how long until they're in new Apple machines? Does anyone know the historical delay? a week? a month?
Why do you want the iMac to remain dual-core? Don't you want Apple to compete?
You can buy Quad Pc's for the price of Mac Mini's these days.
Sometimes I think some Apple Fans are true robots.
I own a Mac Pro, these chips are not designated for Mac Pro (workstation), they are meant for desktops. If Apple can shoehorn a processor in a MacBook Air, they can shoehorn a Quad in an iMac. (Just makes you wonder, how did they get a G5 in there before?).
The problem with this "comparison" is that the Mac Pro is the wrong solution for "I want a quality quad core mini-tower".
So, who really cares what the price is for Dell's wrong solution and how it compares to Apple's wrong solution?
Therefore, it is valid to compare the performance and price of Dell's right solution to Apple's only offering - even if the machines are in different price and performance classes.
Interesting, since one of the SPEC FP tests is H264 video encoding.
I agree that benchmarks are merely estimates, and SPEC numbers are the averages of a quite a few different applications. If you only run one application - then that app is the only benchmark of interest. If you do lots of different things, then a benchmark suite like SPEC is a useful piece of information to consider.
You'll note that I used wiggle-phrases like "can beat" rather than "is faster".
On the H264 tests, the Xeon 2.8 8 core system got a score of 235, and the i7-940 (2.93) got 155. That's 2/3 performance for 1/3 of the price.
The single-stream H264 performance is 33.9 for the i7-940, and 31.0 for the Xeon 2.8. The 8-core Xeon got 95% of expected performance (235/(8*31)), the 4-core Core i7 got 1.15% of expectations (155/(4*33.9)).
The Mini is
- 2.0GHz Intel Core 2 Duo
- 1GB memory 1GB memory
- 120GB hard drive
for $799.
That's funny, you could say the same about a lot of things. The final third is always going to be the most expensive.
No, as I said a few have peaked over. In fact, I purposefully omitted 3,06 Ghz C2D and the 3,8Ghz P4. My main reasons are:
1. The 3,06Ghz C2D is only barely past the 3Ghz mark. Even the 3,2 Ghz Xeon isn't way beyond 3 Ghz (only looking at clock frequency). Compare this to how each new Intel moniker up until the P4 increased clock-speed by more than a factor of two (P2 maxed out at 400Mhz and P3 at 1000Ghz if I'm not mistaken).
2. The 3,8Ghz P4 was a dead end. If I recall correctly the TDP (i.e. "lightbulb factor") was through the roof and (feel free to prove me wrong by presenting some kind of benchmark) the performance improvement over previous processors of the same line was far less than the clock frequency boost would indicate.
Of course processor performance has increased several-fold the last years without increasing clock frequency. My point (being one of pure novelty) was that we (users of Intel desktop processors) jumped back from 3,8 Ghz to 1,66 Ghz to go dual-core (and change tech). And now that we're at about 3Ghz again we most likely will jump back to about 2Ghz to go Quad Core. Also, it's not impossible (pure speculation though) that Xeon processors might go to 2,6 Ghz Octo Core before going to 4Ghz Quad Core.
Also, I'm not sure that saying that Intel regressed to Pentium III to launch Core is a correct analogy (if you have a good source I'd love to read it though) since the Core line is more closely related to mobile processors, such as Centrino(?) which I guess might be derived from the P3, but it's a bit far. (this is moving off the boundry of my knowledge)
This reminds me of the yellow paper that was on top when I opened C&C Tiberium Sun many years ago (after paying several months worth of my then allowance for it). It began something like this:
"YOU DO NOT OWN THIS PRODUCT, THE SOFTWARE, NOR THE MEDIA ON WHICH IT IS CONTAINED..."
Wonder if Apple will start with:
"Engineered AND OWNED by Apple in California"
Buy 3 Dell's, and get twice the performance for the same price as one Mac Pro.
For a render farm, it would be silly to buy 8-core Xeons instead of i7 quads....It would be interesting to benchmark a real multi-threaded video encoder on the two systems.
The SPEC tests are done using 8 single-threaded copies of the program simultaneously.
Depending on the ability of the encoder to use more than four cores, the i7 might do better against the Xeons in real life than the SPEC results show.
The Mini is
- 2.0GHz Intel Core 2 Duo
- 1GB memory 1GB memory
- 120GB hard drive
for $799.
Point proven. Makes it hard to justify a budget desktop Mac purchase with that. Don't throw OS X at me, I'm sorry, look at the friggin photo and tell me I'm wrong.
To bad Apple won't announce that they were going to release computers with quality hardware besides the CPU. CPU means nothing when everything else is as cheap as they come.
Sorry after comparing the parts and build quality between my MBP and my old powerbook I am a little irritated at the new lack of quality on their hardware.
Sorry after comparing the parts and build quality between my MBP and my old powerbook I am a little irritated at the new lack of quality on their hardware.
Quad core with Blu-ray for the price of a Mini....
That is a hard claim to justify based on what we know.
If the pricing holds similar to what we're looking at here, then you'd need twice as many smaller, cheaper boxes to match the power of Gainestown. Bloomfields would be 2/3 the price overall.
If you need ECC memory, or huge memories, or for some reason twice as many boxes isn't workable - then Gainestown might be the right answer.
Myself, I'd wait for SuperMicro Gainestown Twins, though. The Mac Pro maxi-towers are way too humongous for my farm. Why buy those huge boxes when I can get 16 cores (32 threads) in a 1U box? (Mac Pros would be 16 cores in 12U - 12 times bigger)
http://supermicro.com/products/system/1U/1025/SYS-1025TC-T.cfm