31–40 of 57 posts
The speedup is lacking. 12 cores is only 2x faster than 4 cores from 2010. I wonder how the GPU factors into these benchmarks.
Geekbench does not measure GPU performance at all AFAIK.
Nor does it measure I/O. I'm interested to see how that PCIe SSD performs.
Earlier quoted context omitted.
Geekbench does not measure GPU performance at all AFAIK.
> Geekbench does not measure GPU performance at all AFAIK. Nor does it measure I/O. I'm interested to see how that PCIe SSD performs.
Earlier quoted context omitted.
> Geekbench does not measure GPU performance at all AFAIK. Nor does it measure I/O. I'm interested to see how that PCIe SSD performs.
PCIe SSDs are a terrible disappointment.I was able to achive the same performance as ' http://www.newegg.com/Product/Product.aspx?Item=N82E16820227... with 3 Samsung RAID0 drives.
Because that's what Apple says the 2013 Mac Pro SSD is capable of.
Earlier quoted context omitted.
> Geekbench does not measure GPU performance at all AFAIK. Nor does it measure I/O. I'm interested to see how that PCIe SSD performs.
PCIe SSDs are a terrible disappointment.I was able to achive the same performance as ' http://www.newegg.com/Product/Product.aspx?Item=N82E16820227... with 3 Samsung RAID0 drives.
But I don't understand. That card alone is $3,500.
Here are two links which come to a different conclusion from you.
http://www.barefeats.com/mba13a.html
http://www.macrumors.com/2013/06/11/macbook-airs-pcie-based-...
The speedup is lacking. 12 cores is only 2x faster than 4 cores from 2010. I wonder how the GPU factors into these benchmarks.
Actually, looking at disassembly of Geekbench, they barely even use SSE vectors, so it’s really nearly 3/4 of the floating-point performance that goes unmeasured by the benchmark.
One could perhaps argue in the defense of the benchmark that many consumer applications don’t take advantage of these ISA extensions, but you definitely cannot argue that pro apps don’t use them; this alone makes Geekbench (and in fairness, most benchmarks, which tend to be pretty naive) extremely misleading for evaluation of “pro” hardware. Ultimately there’s only one benchmark that’s useful for such evaluation: actually exercising the workload that you intend to run on the machine.
Geekbench also doesn’t use available libraries that offer high-performance implementations of the computations in question (of which any serious application developer would avail themselves). Looking at numbers for released hardware, the “LU decomposition” results in Geekbench make clear that they are either using far too small of a problem size or using a naive C implementation of the operation and measuring the performance of whatever code their compiler generated. If they wanted to show the performance of the system, they would use one of the available library implementations.
Such benchmarks tell you something, but it isn’t really how good hardware is; it’s closer to a measure of how well a compiler can salvage naive C code when it isn’t allowed to take advantage of the available hardware features. I’m not sure what that’s representative of, but I know that it’s not representative of professional workloads.
As someone who's never owned a mac pro can someone give an indication of how often "power" users are upgrading their CPU. When I was a windows desktop user (up until about 8 years ago) I found that everytime i wanted upgrade the internals the CPU socket had changed along with the Mobo chipset necessitating a upgrade of not just CPU, but Mobo and memory. I can understand swapping out gfx cards fairly regularly but are…
As someone who's never owned a mac pro can someone give an indication of how often "power" users are upgrading their CPU. When I was a windows desktop user (up until about 8 years ago) I found that everytime i wanted upgrade the internals the CPU socket had changed along with the Mobo chipset necessitating a upgrade of not just CPU, but Mobo and memory. I can understand swapping out gfx cards fairly regularly but are…
I've used both Mac Pro and HP Z800 workstations, and currently use a HP Z800 that is probably 3 or 4 years old, but also have used a Mac Pro as recently as 2 years ago.
I've never upgraded a CPU in a workstation.
The CPUs I buy are chosen for performance at a sane (relative) price. I'm currently running a couple of Xeon X5560 CPUs, I think they cost £1,600 each at the time.
I also buy as much RAM as is reasonable for the box, I started this Z800 on 24GB ECC RAM which was dirt cheap (relatively - as many slots on the motherboard meant I could get lots of the smaller sizes which were cheap). When the price later dropped on the larger modules I upped it to 192GB RAM.
As for graphics card... now that is something I upgrade every 2 years or so. I started on a Quadro FX1800 and am now on a Quadro 6000.
Aren't the names of these things wonderful? You know you get more when the number is HUGE!
I think this is basically the norm: Upgrade RAM, upgrade GPU... leave CPU until you replace the whole box.
PS: And whilst I'm here, the Mac Pro is a good design... but for the wrong product. I think this is a new design that is the right design with airflow, cooling and silence in mind and would suit desktop users well... but desktop requirements are not workstation requirements and not being able to replace GPUs is a non-starter. I'll skip this design of Mac Pro entirely, or at least until a very wide range of upgradeable GPUs are available. The RAM is also going to be costly, given the few slots available and the higher price for denser modules. I love the design, but this is the wrong class of computer for the constraints that come from the execution of that design.
The speedup is lacking. 12 cores is only 2x faster than 4 cores from 2010. I wonder how the GPU factors into these benchmarks.
Geekbench doesn’t use the GPU; Geekbench also doesn’t use AVX, so half of the floating-point performance of recent cores is left on the table. Actually, looking at disassembly of Geekbench, they barely even use SSE vectors, so it’s really nearly 3/4 of the floating-point performance that goes unmeasured by the benchmark. One could perhaps argue in the defense of the benchmark that many consumer applications don’t tak…
Earlier quoted context omitted.
I don't think performance numbers have been a good way to distinguish between mid-range and high-end for a while now. Especially for users who are looking for a single machine that covers a range of uses including games . ECC memory is a high-end feature. As is the specific choice of GPUs. Arguably, the fast PCIe SSD counts, too, although I suspect that won't last -- everyone benefits from I/O bandwidth, and there's…
So what makes the Mac Pro a mid-range computer? Bear in mind that it's not a consumer product, it's squarely aimed at audio/video/graphics professionals, even though it'll blow any current gaming rig out of the water with those specs.
It's more a Mac Mini Pro than a Mac Pro.
This is a high end machine. Go compare:
http://www.hp.com/united-states/campaigns/workstations/z820_...
Earlier quoted context omitted.
I don't see how a machine that has almost 1GB/s disk transfer speeds, 60GB/s memory bandwidth and 7 teraflops GPU power (almost double the top of the line geforce) is "mid-range" unless you're from the future.
If you do something that is fairly CPU bound, such as large project C++ compiles for iOS software, it matters, a lot. GPUs are not going to make those compiles any faster, and apple has abandoned distcc. The more CPU cores the better. I'm seriously disappointed that they are not offering a 2 or even 3 socket version of the new mac pro. They're all, fuck you iOS developers, the only people who are practically stuck on…