iMac 2020 fluctuating Geekbench scores?

[filmgirl]

Rob has updated his initial test results for the 10-core, it seems he was also tripped up by the RAM installation order in his first run of benchmarks:

2020 iMac 5K 8-core versus 2017 iMac Pro 8-core

real world speed test results for performance minded Macintosh users

barefeats.com

Seems like a pretty convincing lead for the 10-core 5700XT vs. a 2017 8-core iMac Pro and the 8-core 2020 iMac.

I'm feeling sorely tempted by it. My issue (versus offloading almost twice as much for a 7,1 Mac Pro) remains I/O for video work. With only two Thunderbolt 3 ports (sharing a single bus), I have 2x Radeon VII eGPUs, external NVME mass storage and a Decklink Video card to somehow connect up to it (moving them all over from my 5,1 cMP). Which is all rather messier than I'd like it to be.

Does anyone know of a NVME-based NAS (or similar external box) that connects via 10Gbps Ethernet?

QNAP and Synology both have models that support both — Synology even has a combo 10GbE/M.2 adaptor that’ll work with higher-end older models if you’re looking to save some cash/get more bays. QNAP also has a line of TB3 NAS' but be aware that for TB3 to operate faster then 10GbE, the processor needs to be of a certain class — which can make some of the options a bit misleading. I’ve used Synology devices almost exclusively for years and love the product but my experience with QNAP has been great too.

 

[Grumply]

QNAP and Synology both have models that support both — Synology even has a combo 10GbE/M.2 adaptor that’ll work with higher-end older models if you’re looking to save some cash/get more bays. QNAP also has a line of TB3 NAS' but be aware that for TB3 to operate faster then 10GbE, the processor needs to be of a certain class — which can make some of the options a bit misleading. I’ve used Synology devices almost exclusively for years and love the product but my experience with QNAP has been great too.

Do you mean the NVME add-in cards to regular 3.5" HDD NAS boxes? Or are there boxes that allow for a decent number of NVMEs to be installed? I'm pretty keen to completely remove spinning drives from my "online" storage, and just have 8-10TB of NVME for any active projects.

My thinking is actually to avoid using either of the TB3 ports for connecting any storage (using the 10Gb ethernet instead) so that the TB3 ports can both be put towards eGPU/Video I/O cards (to maximise use of the available bandwidth for additional image processing via the eGPUs)

 

[filmgirl]

Do you mean the NVME add-in cards to regular 3.5" HDD NAS boxes? Or are there boxes that allow for a decent number of NVMEs to be installed? I'm pretty keen to completely remove spinning drives from my "online" storage, and just have 8-10TB of NVME for any active projects.

My thinking is actually to avoid using either of the TB3 ports for connecting any storage (using the 10Gb ethernet instead) so that the TB3 ports can both be put towards eGPU/Video I/O cards (to maximise use of the available bandwidth for additional image processing via the eGPUs)

Yes — I would use the TB3s for eGPU too, just wanted to offer that there are 10GbE and TB3 NAS options available. So M.2 in most NAS configs are primarily for caching but you can use a 2.5” SSD in all the bays on most modern (last 4 years or so) NAS's without a problem.

edit to add: some NAS units have dedicated M.2 slots already, some are part of a pre-installed PCI-e card, and some have the option for either a new or additional M.2 slots via an expansion card.

 

[Grumply]

Yes — I would use the TB3s for eGPU too, just wanted to offer that there are 10GbE and TB3 NAS options available. So M.2 in most NAS configs are primarily for caching but you can use a 2.5” SSD in all the bays on most modern (last 4 years or so) NAS's without a problem.

Thanks, sounds like they're probably the only option if keeping the TB3 free for eGPU use is the goal 👍

 

[MauriceG]

I'm deep in the decision process, trying to decide between the i7 and the i9, and also trying to decide upon nano-texture or not.
So I've gone through just about every single review video and post on this forum with benchmarks etc. of the new iMac. But the Geekbench scores seems to vary quite a lot - More than I recall seeing with previous computers.
One person with the i7 gets a score just below 9000 while another person with the same config doesn't get much more than 7500 - Thats a huge difference in actual performance.
Browsing the Geekbench results also shows anything from 6486 to just above 9184 for the i7 3.8 Ghz. Since the i7 version can only be bought in a few different configs, none of which should have much influence on the single-core/multi-core score, I would expect the score to be within 2-300 points in most tests. Can anybody share some insights on this?

Getting an i7 with gives a score of close to 9000 would be nice, but I would be deeply disappointed of getting one that only gives 7-7500.

I haven’t read all messages in this thread, so sorry if my view at the matter doesn’t add value.
During CPU benchmarking the I7 and I9 start to throttle (meaning: the CPU die heats up to 100 degrees Celcius, at that point the CPU will lower the processor frequency to stay below 100, see Intel datasheets). The root cause is that the cooling capacity of the iMac is limited. The temperature of the incoming air than becomes very important. If the air in your room is 20 degrees you will get significant better results than when the environment temperature is 35 degrees Celsius.
A lower temperature of the incoming air will allow the CPU to settle at a higher frequency (i.e. a higher benchmark score)
p.s. according to Intel both the I7 and I9 dissipate nominal 125 Watts (that’s a lot, likely more than any (non-pro) previous iMac model).

 

[Mike Fisher]

My guess is that having FileVault on causes the fluctuations.

 

[RyanXM]

Rob has updated his initial test results for the 10-core, it seems he was also tripped up by the RAM installation order in his first run of benchmarks:

2020 iMac 5K 8-core versus 2017 iMac Pro 8-core

real world speed test results for performance minded Macintosh users

barefeats.com

Seems like a pretty convincing lead for the 10-core 5700XT vs. a 2017 8-core iMac Pro and the 8-core 2020 iMac.

I'm feeling sorely tempted by it. My issue (versus offloading almost twice as much for a 7,1 Mac Pro) remains I/O for video work. With only two Thunderbolt 3 ports (sharing a single bus), I have 2x Radeon VII eGPUs, external NVME mass storage and a Decklink Video card to somehow connect up to it (moving them all over from my 5,1 cMP). Which is all rather messier than I'd like it to be.

Does anyone know of a NVME-based NAS (or similar external box) that connects via 10Gbps Ethernet?

Rob was kind enough to use my testing for those graphs. I tested 4 different types of RAM and I'm continually testing. I'm hoping that Apple unlocks faster RAM speeds as the 3200 and 2933 that I tested only ran at 2666. The Kingston HyperX Impact DDR4-2666 gave the fastest overall throughput in my testing.

iMac (Retina 5K, 27-inch, 2020) - 10-core, 5700XT 16GB, 128GB Kingston HyperX Impact 2666, 4TB SSD - I used all paid versions of all benchmarking apps.

4 Specs of RAM Used
OEM SK Hynix DDR4-2666 (Slowest overall)
Kingston HyperX Impact DDR4-2666 (Fastest overall)
Kingston HyperX Impact DDR4-2933 (Ran at 2666)
Kingston HyperX Impact DDR4-3200 (Ran at 2666)

As far the ThunderBolt 3 ports and their shared bus, the iMac Pro is going to be better because of having 2 buses shared across 4 ports. If you want to go the way of the Mac Pro (2019), then you'll have to cough up more money, but you can move the Radeon VII GPUs internally.

 

[wilberforce]

Always compare the highest geekbench scores you can find for each CPU (or GPU).
There are many reasons why certain geekbench scores are lower (incorrect RAM installation, background processes, etc), but there is no secret magic a particular user does to make their processor run faster than others.
Make sure to disregard any results that appear to be from Macs, but are actually from Hackingtoshes. (Acidanthera, etc)

 

[RyanXM]

Always compare the highest geekbench scores you can find for each CPU (or GPU).
There are many reasons why certain geekbench scores are lower (incorrect RAM installation, background processes, etc), but there is no secret magic a particular user does to make their processor run faster than others.
Make sure to disregard any results that appear to be from Macs, but are actually from Hackingtoshes. (Acidanthera, etc)

My highest on my 10-core: https://browser.geekbench.com/v5/cpu/3358633

 

[vjson]

My highest on my 10-core: https://browser.geekbench.com/v5/cpu/3358633

I'm curious, what was the ambient temperature in your room for that benchmark, and what make/model of RAM have you installed?

EDIT: I just realized you posted several RAM configurations a few posts up, so I take it your best geekbench scores were achieved with the HyperX 2666Mhz modules. I have ordered 4x16GB of those modules myself, and I'm excited to try them out

 

[wilberforce]

I'm curious, what was the ambient temperature in your room for that benchmark, and what make/model of RAM have you installed?

btw, I tried forcing my fan to run at full blast (using Macs Fan Control), to see if I could boost (cheat?) my Geekbench scores that way. I could not discern any significant change in the scores - if there is any, it is less than the normal run-to-run variations.

I also did extensive tests comparing Crucial and Timetec (Hynix) ram, and could tell no difference at all.

There was a small increase (like 2%) for 64GB (4x16GB) vs 32GB (2x16GB) though. Not sure if this is due to more RAM, or because using all four slots (as some have claimed).

 

[vjson]

btw, I tried forcing my fan to run at full blast (using Macs Fan Control), to see if I could boost (cheat?) my Geekbench scores that way. I could not discern any significant change in the scores - if there is any, it is less than the normal run-to-run variations.

Interesting that fan at full blast made no noticeable difference in the scores, since many people are reporting that the iMac is having trouble cooling the i9, resulting in lower sustained boost-speeds compared to the i7. If such is the case, one would think that running the fan at full speed for the i9 would help boost benchmark scores. But perhaps the geekbench test is too short to cause any kind of thermal throttling of the CPU speed.

 

[wilberforce]

Interesting that fan at full blast made no noticeable difference in the scores, since many people are reporting that the iMac is having trouble cooling the i9, resulting in lower sustained boost-speeds compared to the i7. If such is the case, one would think that running the fan at full speed for the i9 would help boost benchmark scores. But perhaps the geekbench test is too short to cause any kind of thermal throttling of the CPU speed.

Mine is the i7, so I would be interested if someone wants to try with their i9

 

[vjson]

Mine is the i7, so I would be interested if someone wants to try with their i9

My bad, got yours mixed up with RyanMX who has the i9. My i9 BTO is ready to be picked up at an Apple store near me, and I'll have it up and running this Friday. I'll definitely run some benchmarks and see if blasting the fan makes any difference for geekbench or cinebench, and I'll report back here.

 

[RyanXM]

I'm curious, what was the ambient temperature in your room for that benchmark, and what make/model of RAM have you installed?

EDIT: I just realized you posted several RAM configurations a few posts up, so I take it your best geekbench scores were achieved with the HyperX 2666Mhz modules. I have ordered 4x16GB of those modules myself, and I'm excited to try them out

This was with the HyperX 2666 and it was run in a room with ambient temps at 71. Fresh install with nothing running in the background and no iCloud account signed in. I also do an SMC and PRAM reset before running any benchmarks. Having swapped the RAM multiple times that day, I found that doing those two resets after a RAM change, it seems to help the scores.

 

[Phil B]

Thanks! I make multitrack music in Reason Studio so the slight single core advantage of i7 is desirable. But I’m conflicted because I would also prefer the i9 10-core multicore speed advantage for X-plane, Final Cut Pro X, and Blender 3D renderings. Any musicians out there in the same boat trying to decide which processor?
[automerge]1597675336[/automerge]

Hi Wendylou, I'm close to replace my old iMac soon, too (was always struggling wait maybe till end of march but anyway nothings points to any updates about the 27" iMac with an M1. And especially speaking about music nothings nearly ready now. So, how was your decision in the end? especially in these mixed world od music, video and 3D (same for me, Logic with lots of Kontakt, and Spitfire BBC SO, video and animation nearly complete Adobe CC (mainly AE, PS, AI, INDD and Lightroom and Blender on the 3D side of daily work and life ;-)

Would really appreciate your experience after a couple of months. Cheers and thanks,
Phil

 

[wendylou]

Hi Wendylou, I'm close to replace my old iMac soon, too (was always struggling wait maybe till end of march but anyway nothings points to any updates about the 27" iMac with an M1. And especially speaking about music nothings nearly ready now. So, how was your decision in the end? especially in these mixed world od music, video and 3D (same for me, Logic with lots of Kontakt, and Spitfire BBC SO, video and animation nearly complete Adobe CC (mainly AE, PS, AI, INDD and Lightroom and Blender on the 3D side of daily work and life ;-)

Would really appreciate your experience after a couple of months. Cheers and thanks,
Phil

I am very pleased with my powerful Intel iMac 2020. I, too, have been monitoring the M1 news and figured it would be a while before a powerful M1 iMac is released. I've also heard the current M1 Mac Mini, as amazing as it is, just isn't quite there yet in terms of comparable power. I'll keep my eyes on any forthcoming M1 iMac that could replace my 2020 iMac, but so far, the Intel iMac is a good investment. It is great for music production with many plugins (I rock Reason Suite); it's fantastic for 4K video editing including HDR Dolby Vision; Blender exports are amazing, and X-Plane 11.5 does not stutter with Orthos, custom scenery, and airports, etc. (although the framerates can slow down if the 3rd party airports and surrounding area are top-heavy with too many 3D objects, that is a challenge to any computer). And note that Apple no longer will offer the iMac Pro because the iMac 2020 can be configured to be more powerful. Anyone on the fence could feel safe getting a new iMac configured to their needs. Put in your own RAM. 64GB is great. I also got a 2TB internal SSD, and the top of the line 16GB graphics option. If they release an M1 iMac that makes me want one, I'm sure I can sell this to someone who wants a machine that can run both MacOS and Windows. I got the 10GB Ethernet, too, to make it more appealing for a resale. I just had the feeling a comparable M1 iMac would take a while to be released, so I feel like I finally did something smart getting the iMac 2020.

 

[Phil B]

I am very pleased with my powerful Intel iMac 2020. I, too, have been monitoring the M1 news and figured it would be a while before a powerful M1 iMac is released. I've also heard the current M1 Mac Mini, as amazing as it is, just isn't quite there yet in terms of comparable power. I'll keep my eyes on any forthcoming M1 iMac that could replace my 2020 iMac, but so far, the Intel iMac is a good investment. It is great for music production with many plugins (I rock Reason Suite); it's fantastic for 4K video editing including HDR Dolby Vision; Blender exports are amazing, and X-Plane 11.5 does not stutter with Orthos, custom scenery, and airports, etc. (although the framerates can slow down if the 3rd party airports and surrounding area are top-heavy with too many 3D objects, that is a challenge to any computer). And note that Apple no longer will offer the iMac Pro because the iMac 2020 can be configured to be more powerful. Anyone on the fence could feel safe getting a new iMac configured to their needs. Put in your own RAM. 64GB is great. I also got a 2TB internal SSD, and the top of the line 16GB graphics option. If they release an M1 iMac that makes me want one, I'm sure I can sell this to someone who wants a machine that can run both MacOS and Windows. I got the 10GB Ethernet, too, to make it more appealing for a resale. I just had the feeling a comparable M1 iMac would take a while to be released, so I feel like I finally did something smart getting the iMac 2020.

Hey, thanks for your detailed answer, that sounds amazing, I'm even more in the mood to order asap hehehe.You went for the i9-10core or i7-8core?

And another question, I heard of some guys having decent issues (especially with the 5700XT) in combination with Adobe PP and other Adobe CC apps (but really horribly reported about Premiere, being nearly unusable)

Do you use any Adobe Stuff? I'm not a big fan of PP, and anyway thinking about going back to FCP or give DaVinci a try. PS, AI, After Effects and some INDD are daily bread and butter for me as well (although I would be really surprised if those fail somehow)

Cheers, Phil

 

[wendylou]

Hey, thanks for your detailed answer, that sounds amazing, I'm even more in the mood to order asap hehehe.You went for the i9-10core or i7-8core?

And another question, I heard of some guys having decent issues (especially with the 5700XT) in combination with Adobe PP and other Adobe CC apps (but really horribly reported about Premiere, being nearly unusable)

Do you use any Adobe Stuff? I'm not a big fan of PP, and anyway thinking about going back to FCP or give DaVinci a try. PS, AI, After Effects and some INDD are daily bread and butter for me as well (although I would be really surprised if those fail somehow)

Cheers, Ph

 

[wendylou]

I bought the i7 8-core after Max Tech suggested that was the one to get after comparison testing. I no longer use Adobe products as I'm not into the whole SAAS rental software model. For video, I used FCPX. I have the 5700XT with 16GB and it is a workhorse for so many things.

 

[Sonicjunki]

New 3.8 i7 owner here, base configuration with 1TB SSD. Have questions for RyanXM on faster ram modules 2933/3200mhz. Wondering if you have tried any other brands and gotten faster speeds than 2667. I guess all those test results of 2933/3200 are from hackentoshes. Shopping for 32GB would like to find the fastest to help my dismal test scores 7387 with apple stock 8GB. Thanks in advance.