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.
iMac 5K GPU Performance
- Thread starter robertojorge
- Start date
- Sort by reaction score
4820k, 4930k, 4960x,5820k,5930k,5960x.
Reference: http://ark.intel.com/products/family/79318/Intel-High-End-Desktop-Processors#@Desktop
The Reason Apple can't use these because of such a high TDP, too small of space to house a 100+ watt TDP CPU. Also, release date I guess
I see what you're sayin', but I was thinking more about the LGA 1150 variety.
I ran a few benchmarks and the 4790K is a pretty low performer over all. Even compared to my 3840QM (labeled "this computer") (non throttling version hehe). I also compared a few other CPU's including a few 6 cores and some slower quad cores. (the xeon is the 6 mac pro)
The software that you're using is not something that I would recommend.
Most of the CPUs that you have selected are 6 core enthusiast line CPUs that are far more expensive than the 4790k. Some of them are 140W and there is no way that can be cooled in an iMac. I am not sure what you were expecting to see when pitting a 6 core vs a 4 core in a multi-threaded benchmark?
Can you see anything faster that could fit into the iMac?
3840QM - 9000 points
http://www.cpubenchmark.net/cpu.php?cpu=Intel+Core+i7-3840QM+@+2.80GHz&id=900
Intel have made incredible strides in the mobile space CPUs in recent years and have let the desktop market stagnate a bit, because they have no competition from AMD any more. We should petition Intel to get themselves sorted out, if you want to see bigger gains in the desktop space.
Last edited:
I look at stock CPU's not overclocked CPUs. Also that site your using is the same site that makes the bench marking program I just used. PassMark Performance Test V8.
Actually I lied, my 3840QM is overclockd to 4.1Ghz. The 4790K I just looked at is not overclocked.
I thought CPU Mark was the old version? Oh well, let's find something else then without over clocked results.
I'm not trying to start a fight. I am just stating facts and I love going back and forth to learn.
I know benchmarks arn't everything but LuxMark does a pretty good job of showing actual OpenCL performance. Someone run LuxMark on their 4790K iMac (just the CPU bench)
Just CPU:
That comes pretty close to a 4790K.
Maybe all this is for nothing.. I was just trying to state that Apple could have put a 6 Core in that machine. Why not? Probably because of the obvious, the Mac Pro.
I think I will get back on topic though given this thread is about GPU performance hehe
Last edited:
No worries at allI'm not trying to start a fight. I am just stating facts and I love going back and forth to learn.
I know benchmarks arn't everything but LuxMark does a pretty good job of showing actual OpenCL performance. Someone run LuxMark on their 4790K iMac (just the CPU bench)
Just CPU:
Image
That comes pretty close to a 4790K.
Maybe all this is for nothing.. I was just trying to state that Apple could have put a 6 Core in that machine. Why not? Probably because of the obvious, the Mac Pro.
TDP would be the biggest issue when it comes to shoving a 6 score into the iMac. I don't think Apple are too concerned about cannibalising their sales. From a business perspective, I don't really care if a customer buys a $4000 iMac or a $4000 Mac Pro.
Your original score for the 4790K still makes no sense to me.
No worries at all
TDP would be the biggest issue when it comes to shoving a 6 score into the iMac. I don't think Apple are too concerned about cannibalising their sales. From a business perspective, I don't really care if a customer buys a $4000 iMac or a $4000 Mac Pro.
Your original score for the 4790K still makes no sense to me.
I think you're right about the sales thing lol
As for the benchmark, I looked at the guys uploaded config file he sent to passmark and saw his turbo boost numbers and noticed that they were set to stock so I assume it's stock performance.
I think you're right about the sales thing lol
As for the benchmark, I looked at the guys uploaded config file he sent to passmark and saw his turbo boost numbers and noticed that they were set to stock so I assume it's stock performance.
I agree that these synthetic benchmarks don't always mean much. They usually show that this processor can score xxx in this particular benchmark, especially as most of them don't stress a CPU enough to invoke thermal throttling, but have you seen the processor scores on geekbench? The i7 4790k kills all in the single core tests and does pretty well for itself in the multi core.
http://browser.primatelabs.com/processor-benchmarks
(this is for the standard CPU only, no over clock)
This is why I was questioning the score you post and thought you were using an odd version of the software.
Yeah and Geekbench for me is 3600 single core peak. That seems to line up since I am only 300Mhz slower at 1 core than a 4790K.
I am happy I can compete with a 4790K no doubt lol
Good old anandtech to the rescue with a slightly more real world test to give some perspective on these "scores"
For what it is, the 4790K is a powerhouse in my opinion.
The first bench is using 1 core. I mean, no doubt having one core screaming at 4.4Ghz makes some impressive single core scores.
Handbrake always uses all available cores, even when ripping DVD quality sources. The point is that in that particular workflow, the 4790K works well. You're not going to find anything better at the price point and under 90W TDP.
The iMac keeps closing the gap between it and my 2009 8c with 5870. I really don't want to put a GPU in with external PSU and all that. Would be nice to start fresh and maybe take a performance hit but not noticeably. I think the better I/O and internal storage will help bridge the performance gap even more.
Hmm.
Hmm.
No worries at all
TDP would be the biggest issue when it comes to shoving a 6 score into the iMac. I don't think Apple are too concerned about cannibalising their sales. From a business perspective, I don't really care if a customer buys a $4000 iMac or a $4000 Mac Pro.
Your original score for the 4790K still makes no sense to me.
I think the die size for the 6 core would be larger and it couldn't cope with Haswell though broadwell possibly. Thing is that iMac and MacBook Pro are built on a compromise of balancing TDP with the slim chassis and mobile chipsets, whereas the Mac Pro has far more headroom cos the towers uses PCIe Cards in a very large case with a lot of airflow and the TDP of the Mac Pro 6,1 design with the thermal core is on another planet - for me by far the most impressive thing about that little black can.
Indeed. Personally I did not believe the claims of it being quiet after my experiences with MBPs and Mac Minis, but I am still amazed after 7 months of use.
Not to keep commenting on this but here is a benchmark that someone is running on a Bootcamped riMac.
Yeah.... maybe throttling is going on. Again this is compared to my 3840QM @ 4.1Ghz
Yeah.... maybe throttling is going on. Again this is compared to my 3840QM @ 4.1Ghz
Most of my clients for these have been windows based, running for example 6 hour 3dsmax renders in win 8.1 and they can't hear them while the hp z series boxes nearby sound like chinooks. They are now converts and for single cpu workstations are switching to the 6,1.
They are laughably ludicrously quiet and having seen an hour of bf4 using crossfire d700's in bootcamp during lunchtime with high settings is an aural experience to not behold. A gentle breath sound instead of an absolute din you always should hear. Only serious water cooled kit have got close and that's still nowhere near the lack of noise.
I would love one - but not to replace my cheese-grater, that's far too useful for my needs as a dual boot SATA sledded beast for fixing Mac and PC gear. It would be my transcoding mute and gaming rig!
It's apple's thermal paste lottery, some will throttle some won't. Had it with all the mbp retinas and the iMac's and I have no doubt these 5k models will vary wildly too. A repaste and polish they throttle far less. Having said that I've even got the Mac Pro 6,1 running 5-6c cooler upgrading from the 4 core to the unauthorised 3.3 8 core. Apple consistently paste as well as a one eyed decorator with glaucoma
Last edited:
Folks,
I am very excited to be able to present to you some Windows benchmarks for the R9 M295X!
Seeing the earlier post about Passmark I wondered if I could pull some iMac benchmarks, and I could. There wasn't any for the R9 M290X so that's a shame. Now whether you value Passmark as a graphics benchmark is up to you, but at least it's something and more info than we had before.
The M295X was running Windows 8.1. The other results are iMacs from the appropriate year running Windows 8. 'This computer' is my laptop, a mid-2012 rMBP running Windows 7.
I have no idea what is up with the 2D results, driver issues?
I am very excited to be able to present to you some Windows benchmarks for the R9 M295X!
Seeing the earlier post about Passmark I wondered if I could pull some iMac benchmarks, and I could. There wasn't any for the R9 M290X so that's a shame. Now whether you value Passmark as a graphics benchmark is up to you, but at least it's something and more info than we had before.
The M295X was running Windows 8.1. The other results are iMacs from the appropriate year running Windows 8. 'This computer' is my laptop, a mid-2012 rMBP running Windows 7.
I have no idea what is up with the 2D results, driver issues?
Um, can you give us a larger image? I can't make out anything on those graphs.