If there are no thermal constraints the i9 10900K will outperform the i7 10700K in single core too because of the higher boost clocks. 5.3GHz vs 5.2GHz, but inside the iMac, the boost clocks for both are capped at 5.0GHz. With the thermal system and constraints the i7 10700K will likely beat the i9 10910 in single-core.
You see this play out with the 10600K which has a base clock of 4.1GHz over the i7 or i9 in more constrained systems where the boost clocks cannot be sustained.
No thermal constraints when running 1 core 100%. According to datasheets at ark.intel.com the max. clock speed for the i7 is 5.1 GHz en 5.0 for the i9. Apple hasn’t capped the clock speed; they do use tailored versions of the CPUs.
The i9/5700XT combo is slowly starting to scare me lol but according to Max Tech it shouldn't be too bad.
I agree you absolutely do not need a top spec machine for LR and PS. I went from a 2014 5k iMac to a 2020 5k iMac (both i7), and see only slight improvement in responsiveness for photo-editing (aside from batch exporting or importing) - on both machines the responsiveness is virtually immediate to any adjustment, with little to no perceptible lag.I'd say - save your money. I'm a professional photographer and videographer. I edit RAW images in LR and PS as well as do 1080 and 4K video editing in Premiere. I'm on a 2013 iMac i5 I bought used for $650 on eBay last year. I've got a 512 SSD and believe it or not only 8GB of RAM. It handles everything I throw at it. I can edit 4K footage no problem... LR could be a bit faster, but it works and I SAVED a ton buying it.
8thNote - would you mind expanding on this a little please? How much better was 4 x 16 compared to 2 x 32 and how did you test them/tell the difference?
About to order 64GB and wondering which to get. I'd get 2 x 32GB if no real difference as that would allow possibility for upgrade later but if 4 x 16 is noticeably better I might do that - also wondering if 4 x 16GB might run a little cooler.
There is virtually no difference between 4x16 or 2x32, really.
Fastest two 2020 models are 8th and 11th from the top on the "Multi-Core" list (see screenshot.)
Sorry, but I call BS on this lolI'd say - save your money. I'm a professional photographer and videographer. I edit RAW images in LR and PS as well as do 1080 and 4K video editing in Premiere. I'm on a 2013 iMac i5 I bought used for $650 on eBay last year. I've got a 512 SSD and believe it or not only 8GB of RAM. It handles everything I throw at it. I can edit 4K footage no problem... LR could be a bit faster, but it works and I SAVED a ton buying it.
That is the fastest combo on the 2020 iMac. Period.
All this noise and 300 posts to say that it is not worth the extra $400 is very entertaining. But at the end of the day, the fastest 2020 iMac is the i9/5700XT combo.
Is it true that CPU can be upgraded? And how difficult it is?
Everyone seems to think these are the last Intel iMacs, but I'm not so sure. While I won't be surprised if the next top performing iMacs are using Apple's chips, I think transitioning the entire Mac lineup to ARM could be a slow multi-year process. Just because Apple announced ARM Macs are coming doesn't mean their relationship with Intel is completely over.
As for this year's i7 vs. i9, choosing the 8-core doesn't seem to be compromising much in the way of performance, so I put the money into a GPU upgrade instead.
72GB RAM (adding 2, 32GB OWC RAM to the stock 8)
Not really... here's are my findings (originally posted on another thread):
I gathered 4 pages worth of benchmarks from Geekbench for each of the following: 8-core w/either 5700 (8GB VRAM) or 5700XT (16GB VRAM), 10-core w/either 5700 (8GB VRAM) or 5700XT (16GB VRAM)
Deleted the highest and lowest score and averaged them out:
iMacs with =<64GB system RAM
8C/8GB (9 samples): 46023
8C/16GB (26 samples): 54232*
10C/8GB (2 samples): 43639
10C/16GB (15 samples): 55113
iMacs with >64GB system RAM
8C/8GB (2 samples): 44741
8C/16GB (37 samples): 54776
10C/8GB (3 samples): 46079
10C/16GB (67 samples): 54753
From the numbers, you can see that going from 8GB VRAM to 16GB makes a big difference. Going from 8-core to 10-core has minimal impact and even in some cases a lower score. And seems like going beyond 64GB of RAM provides no benefits at all. So in the end, I went for the sweet spot: 8-core i7/5700XT/64GB*
Errors to take into account: ppl submitting multiple benchmarks skewing the data, RAM installed non-optimally, non-controlled environment
Someone already pointed out your ram misconfiguration but the video card is also not doing anything for you. Photo editing makes very little use of GPUs. 4GB VRAM is more than enough for this kind of work. 4K video is a different story.I just installed two top line I-7 27" iMacs for senior photographers. I didn't even consider the matte screen, and decided that the i7 was a better choice for the price. Neither of these photographers shoot video professionally, so I believe the i7 would be sufficient along with the high end video card and 72GB RAM (adding 2, 32GB OWC RAM to the stock 8). Not 72MB. I haven't heard any complaints yet, they really scream. However, I think there are very good reasons to go for the i9, as it has proven itself as the workhorse of the Intel spread. If I was doing any 4K or uncompressed video editing, I think it would be worth the extra $. My only complaint about the iMacs are the screens, especially for printing. After using internally calibrated high bit processing screens (Eizo and NEC) for the last decade, I can never see gray on a Mac screen... I could definitely see using an Eizo in combination with an iMac 27" but not while my 12-core cMP is alive and well in Catalina, thanks to MacRumors fora.
I put the chips in different configurations and then tested with Geekbench 4 and 5. I liked Geekbench 4 better for testing RAM performance because it would measure RAM bandwidth.
You're welcome to look at my test results: https://browser.geekbench.com/user/338794
Look at the note on each recent test to see my top-to-bottom RAM chip configuration (like 0 32 0 32).
Thanks, this is interesting nice job!! Is there a reason you used Geekbench 4? Geekbench 4 is more dependent on RAM and GPU than Geekbench 5, right?
I am mainly interested in the CPU difference since that will be my bottleneck, it seems that in your results the GPU (although I doubt if it's the VRAM, I guess it's the better chip of the 5700XT but I'm no specialist at all) makes the difference like you said, not the CPU. Would that be different in Geekbench 5? There is a relatively substantial difference between the i7 and i9 in the Mac charts there.
Also, in your results, you use two categories. =<64 GB of RAM and > of RAM. Wouldn't it be better two just choose the two configurations people use the most, like 32GB and 64GB? In your =<64 GB category there are probably people who run their benchmarks with 8GB of RAM before upgrading with aftermarket RAM, and that will definitely be a bottleneck right? And I guess nobody with an i7 or i9 is going to stick to 8GB of RAM.
I was really going for the i9 but after weeks of reading about it I may be losing my religion. I would love somebody from Apple explaining why they messed up that beautiful 10th gen i9 opportunity..
Does anyone here know how to check if you have 10Gb ethernet using "About this Mac"?In a couple weeks (saving my pennies) I plan to order a new 2020 27" iMac to replace my mid-2011 (4 core i5). I have a question/ looking for opinions...
What I know I'll be getting:
- One of the top-tier processors
- 5700 XT 16GB Radeon graphics.
- 8GB memory (and buy additional 64GB at OWC).
- 1TB SSD (99% of my working photos and videos are stored on external drives).
- Keyboard with added number pad
- Magic Mouse (No interest in the Magic Track Pad - I have a 1st gen track pad and never use it)
- 10GBit Ethernet
- Final Cut Pro X
Honestly, I wasn't aware of different geekbench versions. I was just happy to get my hands on some data when I stumbled upon those because it identified the processor, OS version, RAM amount and GPU used. How do you even know it was version 4 or whatever?
I basically split into two RAM groups because RAM amounts were all over the map - 8, 24, 32, 48, 64, 72, 128 and it would have been too complicated, plus I would have lost a lot of data samples and some configs were rare. It was meant to give 30k ft view