Question Would an i7-12700 EVER make more sense than an i5-12600K ?

[My PC Hates Me]

Would an i7-12700 (non-K version, not the F version) ever make more sense than an i5-12600K ???

I know the 12-700 is more expensive than the i5-12600K

Could the price difference be offset by a less expensive motherboard? less expensive cooler / fan(s)?

It would mostly be used for video editing (in DaVinci Resolve) and Photo editing (Lightoom, Photoshop, gimp, darktable). Maybe some esports gaming.

 

[Karadjgne]

DaVinci Resolve will use up every single thread it can get its grubby hands on. Also to consider is the 12700 has more E-cores than a 12600k.

The 12700 is a workhorse cpu, the 12600k is a racehorse, albeit not the fastest.

So yes, the 12700 makes a ton more sense than the 12600k, in every way. Not only from a core perspective, but it's lack of OC means Zero temptation to do anything dumb like overspending on a Z690.

Cooling will be important. Long sessions in DaVinci, pushing every thread, is going to push the heat. Maybe. Using default power limits you'll be using @ 130w ±, but lifting the power limits and setting the cpu free to do as it can will raise the power consumption over 250w, that 280mm/360mm AIO territory, not big-air unless you like wind tunnel noise.

 

[jasonf2]

You are picking up two more performance cores, which will show up as 4 more threads. This will give you 25% more performance cores when utilizing highly parallelized multithreaded applications. Those workloads usually are things like video editing and transcoding depending on how the software is programmed. For video games you are probably not going to really notice a difference as the clocks and ipc are identical at the core level. Is a hundred dollar difference worth it? That is too subjective for me to answer.

 

[My PC Hates Me]

DaVinci Resolve will use up every single thread it can get its grubby hands on. Also to consider is the 12700 has more E-cores than a 12600k.

The 12700 is a workhorse cpu, the 12600k is a racehorse, albeit not the fastest.

So yes, the 12700 makes a ton more sense than the 12600k, in every way. Not only from a core perspective, but it's lack of OC means Zero temptation to do anything dumb like overspending on a Z690.

Cooling will be important. Long sessions in DaVinci, pushing every thread, is going to push the heat. Maybe. Using default power limits you'll be using @ 130w ±, but lifting the power limits and setting the cpu free to do as it can will raise the power consumption over 250w, that 280mm/360mm AIO territory, not big-air unless you like wind tunnel noise.

Thank you for the reply.

Hmmm... I was hoping to avoid going with liquid cooling. Thanks for the heads up about needing an AIO.

 

[My PC Hates Me]

You are picking up two more performance cores, which will show up as 4 more threads. This will give you 25% more performance cores when utilizing highly parallelized multithreaded applications. Those workloads usually are things like video editing and transcoding depending on how the software is programmed. For video games you are probably not going to really notice a difference as the clocks and ipc are identical at the core level. Is a hundred dollar difference worth it? That is too subjective for me to answer.

Thanks for the reply.

I did see a few benchmarks online (which were focused on gaming) and it looks like in terms of fps, they are basically tied. Then again, my main game of choise is Solitaire, so I don't think it is going to matter much to me. (Ok, ok, I confess, I sometimes play Paladins. Please don't laugh.)

I did see a few other benchmarks related to rendering in hevc / havc and it does look like the extra cores / threads of the i7-12700 really pay off (reducing rendering time by about 25%)

I am assuming that SOME operations in Lightroom and Photoshop make use of multiple cores / threads. For instance, on my current i7-6770, when applying a global preset to a bunch of photos in lightroom, it looks like ALL the threads are pegged at 100%. Just guessing here, but assuming things like generating previews and exporting jpg files would be faster.

 

[My PC Hates Me]

I guess my other question is whether I will end up having to get a Z690 motherboard anyway because I heard a rumor that the B660 motherboards won't be compatible with 13th-gen chips when they come out?

(My plan was to wait for 13th gen chips to come out and see if I can find a cheap-ish 12th gen chip. With all the craziness in the world, who knows if that's even gonna work or not.)

 

[Karadjgne]

12600k is 6/12 P-cores, 4 E-cores, 20MB L3 cache.
12700 is 8/16 P-cores, 4 E-cores, 25MB L3 cache.

For gaming, the 12700 generally beats the 12600k, simple matter of Lcache size since both cpus run the same 4.9GHz clocks.

DaVinci will use every available thread, so even just using the P-cores, it'll come out ahead. In any multithreaded workload, the 12700 wins. According to any benchmark available, in blender the 12600k comes in around 88% vs 1200k. That @ 8minutes savings on a 1 hour compile, every time it's used.

And that's at the 65w PL1, not the 125w PL2 which can push all cores at 4.8GHz instead of standard turbo settings.

That's worth the price difference. 12600k is $278, 12700 is $342, $64 difference single cost, and then consider the cost difference between a Z690 and B660/H670 boards.

B660, H670, Z690 are the same socket, same layout, basically the same motherboard ± some bells and whistles and extra components. The difference is in the chipset. I seriously doubt Intel would be stupid enough to not use a B660 and 13th gen cpu, like a 13100 or 13400 or even a 13700. You can stick a 11900 on anything that'll accept a 10 series cpu.

The only caveat to that is the board vendors themselves. They might decide that their budget boards VRM's are not suitable for the 125w cpus, so might preclude the 125w capability of the 12700 but be absolutely fine with the 65w PL1 values. I'd assume that'd follow for the 13th gen cpus too.

 

[Nighthawk117]

Would an i7-12700 (non-K version, not the F version) ever make more sense than an i5-12600K ???

I know the 12-700 is more expensive than the i5-12600K

Could the price difference be offset by a less expensive motherboard? less expensive cooler / fan(s)?

It would mostly be used for video editing (in DaVinci Resolve) and Photo editing (Lightoom, Photoshop, gimp, darktable). Maybe some esports gaming.

The short answer is yes it would, it has 8 performance cores instead of 6. It would depend on what you meant by less expensive motherboard. You won't get away with less cooling though. Both chips would be very capable at the workloads listed above.

 

[My PC Hates Me]

The short answer is yes it would, it has 8 performance cores instead of 6. It would depend on what you meant by less expensive motherboard. You won't get away with less cooling though. Both chips would be very capable at the workloads listed above.

Thanks for the reply!

Regarding your question, "It would depend on what you meant by less expensive motherboard."

I figured that since the i7-12700 can't be overclocked, that I could go with a (cheaper) B660 motherboard instead of a (more expensive) z690 board.

But I don't really know that much about motherboards so don't know if that will be a mistake or not. and as mentioned above, I heard "somewhere" that when 13th-gen chips are released, they will only work on z690 boards and not on b660 boards. But I could easily have misunderstood that.

 

[My PC Hates Me]

12600k is 6/12 P-cores, 4 E-cores, 20MB L3 cache.
12700 is 8/16 P-cores, 4 E-cores, 25MB L3 cache.

For gaming, the 12700 generally beats the 12600k, simple matter of Lcache size since both cpus run the same 4.9GHz clocks.

DaVinci will use every available thread, so even just using the P-cores, it'll come out ahead. In any multithreaded workload, the 12700 wins. According to any benchmark available, in blender the 12600k comes in around 88% vs 1200k. That @ 8minutes savings on a 1 hour compile, every time it's used.

And that's at the 65w PL1, not the 125w PL2 which can push all cores at 4.8GHz instead of standard turbo settings.

That's worth the price difference. 12600k is $278, 12700 is $342, $64 difference single cost, and then consider the cost difference between a Z690 and B660/H670 boards.

B660, H670, Z690 are the same socket, same layout, basically the same motherboard ± some bells and whistles and extra components. The difference is in the chipset. I seriously doubt Intel would be stupid enough to not use a B660 and 13th gen cpu, like a 13100 or 13400 or even a 13700. You can stick a 11900 on anything that'll accept a 10 series cpu.

The only caveat to that is the board vendors themselves. They might decide that their budget boards VRM's are not suitable for the 125w cpus, so might preclude the 125w capability of the 12700 but be absolutely fine with the 65w PL1 values. I'd assume that'd follow for the 13th gen cpus too.

Thanks so much for the elaboration. I really appreciate it.

Yeah, it looks like the 12700 might be the way to go for me as I literally only go on a game-binge like one weekend once every two or three months (which probably explains why I am so bad). Then I literally uninstall for two to three months, reinstall, play for a weekend, uninstall, etc...

 

[Nighthawk117]

Thanks for the reply!

Regarding your question, "It would depend on what you meant by less expensive motherboard."

I figured that since the i7-12700 can't be overclocked, that I could go with a (cheaper) B660 motherboard instead of a (more expensive) z690 board.

But I don't really know that much about motherboards so don't know if that will be a mistake or not. and as mentioned above, I heard "somewhere" that when 13th-gen chips are released, they will only work on z690 boards and not on b660 boards. But I could easily have misunderstood that.

A B660 is perfectly fine yes, however not a cheap B660, you will want one with robust power delivery for an i7 12700. Something like a B660 Mag Mortar (mATX) or B660 Mag Tomahawk (ATX) would be my recommendation. That's probably true about 13th gen and B660, however inter-generational upgrades are almost never worth it IMO. 13th gen has a nice bump in multi threaded performance due to extra E cores but it's a very incremental step from 12th gen.

13th gen is out in October, that's not too far away, so you could always wait for that and get the 13700. If your at all concerned about upgrade paths on the same platform then my two cents would be to simply buy the best CPU you can afford like the i7, that way it will last you longer and you won't have to worry.

 

[Lafong]

I heard "somewhere" that when 13th-gen chips are released, they will only work on z690 boards and not on b660 boards. But I could easily have misunderstood that.

Only announcement I've seen is that MSI has stated that a BIOS upgrade is needed if you want to put a 13th gen CPU on a B660 board:

https://www.msi.com/news/detail/MSI...s-Support-The-Next-Gen-Intel-Processors140590

 

[My PC Hates Me]

A B660 is perfectly fine yes, however not a cheap B660, you will want one with robust power delivery for an i7 12700. Something like a B660 Mag Mortar (mATX) or B660 Mag Tomahawk (ATX) would be my recommendation. That's probably true about 13th gen and B660, however inter-generational upgrades are almost never worth it IMO. 13th gen has a nice bump in multi threaded performance due to extra E cores but it's a very incremental step from 12th gen.

13th gen is out in October, that's not too far away, so you could always wait for that and get the 13700. If your at all concerned about upgrade paths on the same platform then my two cents would be to simply buy the best CPU you can afford like the i7, that way it will last you longer and you won't have to worry.

Thank you for the explanation about the different motherboards, and for the suggestion to wait it out a bit. Much appreciated.

 

[My PC Hates Me]

Only announcement I've seen is that MSI has stated that a BIOS upgrade is needed if you want to put a 13th gen CPU on a B660 board:

https://www.msi.com/news/detail/MSI...s-Support-The-Next-Gen-Intel-Processors140590

Thank you for the link. I appreciate that.

 

[Karadjgne]

Intel has been doing this for years, it's a tick-tock approach. 2nd gen/3rd gen, 4th/5th, 6th/7th, 8th/9th, 10th/11th, and now 12th/13th. Changing sockets is a very big deal, it requires a total revamp of all traces and circuits, so Intel needs the 'tock' time to finalize the next tick in order to get that info to the AIB partners like Asus, Msi, Gigabyte etc, so they can start work on the next tick boards.

13th to 14th gen will be a larger step advance than 12th to 13th gen will be, same as Broadwell to Skylake was a bigger jump compared to Devils Canyon to Broadwell.

I'd not worry about any possible performance loss by going 12th gen vs 13th gen, in games a 10% performance increase is generally 3-5fps and when you are already pushing over 100fps, you aren't going to even see a 10fps increase. Can't tell the difference.

It's like fast boot. That works great for boot time savings when using a Hdd, that's only read/write 180Mb/s, but absolutely pointless for an NVMe that's already read/write 1000Mb/s. You save 1-2 seconds. Wait, did you blink and miss those 2 seconds?

 

[jasonf2]

Thanks for the reply.

I did see a few benchmarks online (which were focused on gaming) and it looks like in terms of fps, they are basically tied. Then again, my main game of choise is Solitaire, so I don't think it is going to matter much to me. (Ok, ok, I confess, I sometimes play Paladins. Please don't laugh.)

I did see a few other benchmarks related to rendering in hevc / havc and it does look like the extra cores / threads of the i7-12700 really pay off (reducing rendering time by about 25%)

I am assuming that SOME operations in Lightroom and Photoshop make use of multiple cores / threads. For instance, on my current i7-6770, when applying a global preset to a bunch of photos in lightroom, it looks like ALL the threads are pegged at 100%. Just guessing here, but assuming things like generating previews and exporting jpg files would be faster.

Sorry for the delay.

As far as Lightroom and Photoshop go they absolutely make use of multiple cores on both CPU and GPU for rendering. However the amount of processing power necessary to render a single image or filter in a timely fashion is easily met with most mainstream setups today with a decent amount of RAM and an SSD. Video, because of the shear volume of frames to be rendered at ever increasing resolution, still can take a significant amount of time to render and transcode. So depending on what you are doing and if you are doing it for profit the difference of waiting a couple of minutes on a transcode (or effect render) or a couple of hours (or even days) pretty quickly justifies the difference between your I5 with integrated graphics and 16 gigs of RAM and the latest Threadripper pro, a stacked set of Quadro video cards and hundreds of gigs of RAM. The same goes for stuff like AI training or anything processing huge amounts of data. As cool at that workstation rig sounds though it really wouldn't be that great for gaming.
For all given purposes you probably cannot tell much of a difference between these processors in Lightroom or Photoshop unless you are running a large batch job. On the other hand if you are working in After Effects or Premier Pro regularly that is where you will see a bump when running final render, adding effects and/or transcode on you finished work. Regardless lots of RAM, a fast SSD and strong video card will be just as important if not more so in those scenarios than a couple of extra P cores.

 

[logainofhades]

I would choose a 12700, with a good enough B660 board, over a 12600k, with a good enough Z690, any day of the week.

 

[Albert.Thomas]

It would mostly be used for video editing (in DaVinci Resolve) and Photo editing (Lightoom, Photoshop, gimp, darktable). Maybe some esports gaming.

In all of those scenarios, the 12700 is a better choice. It has more cores, which will be useful in video editing. It has more L3 cache, which will be more useful in esports.

 

[My PC Hates Me]

Sorry for the delay.

As far as Lightroom and Photoshop go they absolutely make use of multiple cores on both CPU and GPU for rendering. However the amount of processing power necessary to render a single image or filter in a timely fashion is easily met with most mainstream setups today with a decent amount of RAM and an SSD. Video, because of the shear volume of frames to be rendered at ever increasing resolution, still can take a significant amount of time to render and transcode. So depending on what you are doing and if you are doing it for profit the difference of waiting a couple of minutes on a transcode (or effect render) or a couple of hours (or even days) pretty quickly justifies the difference between your I5 with integrated graphics and 16 gigs of RAM and the latest Threadripper pro, a stacked set of Quadro video cards and hundreds of gigs of RAM. The same goes for stuff like AI training or anything processing huge amounts of data. As cool at that workstation rig sounds though it really wouldn't be that great for gaming.
For all given purposes you probably cannot tell much of a difference between these processors in Lightroom or Photoshop unless you are running a large batch job. On the other hand if you are working in After Effects or Premier Pro regularly that is where you will see a bump when running final render, adding effects and/or transcode on you finished work. Regardless lots of RAM, a fast SSD and strong video card will be just as important if not more so in those scenarios than a couple of extra P cores.

Thank you for the reply.

In general, I will need to be doing a total "system" upgrade anyway, since I am still rocking an i7-6700 with the OS and apps on a SATA SSD, and 24GB of DDR 4 memory (2400Mhz??? It's slower memory). My lightroom catalog and RAW / JPG files are on spinning SATA hard disks.

I have an RTX 2060 Super with 8GB of VRAM and I occasionally get out-of-memory errors in Resolve.

Surprisingly, my current system still "works," but it is getting long in the tooth.

So yes, looking to upgrade so I can add more RAM and faster SSDs, and get a better video card than my RTX 2060 Super.

Oh, the other "time suck" when video editing is when I create (lower-resolution) proxy files to edit from. So I am hoping to either 1) avoid having to create proxy files (meaning my system is fast enough to edit / scrub 4K hevc smoothly), or,

2) Create proxy files much faster.

 

[My PC Hates Me]

I would choose a 12700, with a good enough B660 board, over a 12600k, with a good enough Z690, any day of the week.

Thanks for the reply.

But why go with the 12700 and b660 over the 12600K and a z690?

Not saying I disagree with you (in fact, that is the way I am leaning). Just curious as to why YOU would do it.

Thanks in advance.

 

[My PC Hates Me]

In all of those scenarios, the 12700 is a better choice. It has more cores, which will be useful in video editing. It has more L3 cache, which will be more useful in esports.

Thank you for the reply. Although lately, most of my gaming has been limited to wordle... But probably in a month or two I will get the bug to reinstall Paladins again and be a meat shield, as I usually am.

 

[Karadjgne]

A 12600k and Z690 costs roughly the same as a 12700 and B660. There's a performance difference right off the rip, the 12700 beats the 12600k in games and production. The Only benefit of the 12600k is its ability to OC, which is fine for gaming, but introduces the chances of instability, which is Death to production in things like video editing etc, where a single bit-flip can ruin a render, add funky pixels in frames etc. For production, stability is key, and that doesn't happen with OC.

So pay for performance out of the box, or pay for possible speed increases and its inherent hazards. Me personally, I'd not want a glitch happening 3/4 of the way through a 6hr render, forcing a complete start-over.

 

[jasonf2]

Thank you for the reply.

In general, I will need to be doing a total "system" upgrade anyway, since I am still rocking an i7-6700 with the OS and apps on a SATA SSD, and 24GB of DDR 4 memory (2400Mhz??? It's slower memory). My lightroom catalog and RAW / JPG files are on spinning SATA hard disks.

I have an RTX 2060 Super with 8GB of VRAM and I occasionally get out-of-memory errors in Resolve.

Surprisingly, my current system still "works," but it is getting long in the tooth.

So yes, looking to upgrade so I can add more RAM and faster SSDs, and get a better video card than my RTX 2060 Super.

Oh, the other "time suck" when video editing is when I create (lower-resolution) proxy files to edit from. So I am hoping to either 1) avoid having to create proxy files (meaning my system is fast enough to edit / scrub 4K hevc smoothly), or,

2) Create proxy files much faster.

For your described workloads the more cores the better. If you are building a new machine don't overlook a comparable Zen 3 product. Intel is still playing catchup on process node so their per core efficiency is behind AMD. The big little configuration that Intel has put into place gives them the ability to put high clocks on their P cores. This makes them power hungry chips that are good for video game frame rates. This doesn't necessary make them the best for heavily multithreaded workloads though. AMD is still running identical cores. So if I take a 5900X (12core zen 3) and your 12700 the 12700 will out perform in single threaded applications by roughly 20%, but lose in a multi-threaded load by ~5%. The alder lake product is newer and has the more modern PCI 5 standard, but in all fairness nothing mainstream really maxes out multiplexed 4 right now. PCI 5 will allow for faster SSDs, but reasonably priced consumer drives are still a little bit out. Also if you are going to run an alder lake cpu make sure to go to Windows 11. To get the performance out of the big-little you need it's updated task scheduler. Best of luck.

 

[My PC Hates Me]

For your described workloads the more cores the better. If you are building a new machine don't overlook a comparable Zen 3 product. Intel is still playing catchup on process node so their per core efficiency is behind AMD. The big little configuration that Intel has put into place gives them the ability to put high clocks on their P cores. This makes them power hungry chips that are good for video game frame rates. This doesn't necessary make them the best for heavily multithreaded workloads though. AMD is still running identical cores. So if I take a 5900X (12core zen 3) and your 12700 the 12700 will out perform in single threaded applications by roughly 20%, but lose in a multi-threaded load by ~5%. The alder lake product is newer and has the more modern PCI 5 standard, but in all fairness nothing mainstream really maxes out multiplexed 4 right now. PCI 5 will allow for faster SSDs, but reasonably priced consumer drives are still a little bit out. Also if you are going to run an alder lake cpu make sure to go to Windows 11. To get the performance out of the big-little you need it's updated task scheduler. Best of luck.

Thanks so much for the well thought out (and clearly articulated) reply.

There are two points I just want to bring up (mostly to see whether I am over-thinking things)

1) My main reasoning for going with Intel 12th gen over a comparable AMD-based system is that my video program of choice (DaVinci Resolve) works better with Intel (and Nvidia) than with AMD processors or graphics cards. DaVinci Resolve can utilize the iGPU in 11th- and 12th-gen Intel processors for encoding and decoding lots of codecs that it can't do with AMD cpu's. Hence that is why I am leaning HEAVILY toward Alder Lake or even possibly 13th gen intel cpu.

NOTE: I sometimes think about going the "cheap route" and buying a used business desktop machine. There are some i7-10700 (non-K) HP desktops around here for $400 that have typical business specs (500GB nvme and a 1TB spinning HDD and NO graphics card). I am sure it would be a lot faster than my i7-6700 once I throw in my current RTX 2060 super and my 650 Watt Gold EVGA power supply (if it will fit, might need to invest in a hack saw) and be done with it. But I anticipate that the ability to encode / decode video codecs is something I am going to want further down the road.

2) Your comment about Windows 11 has me a bit concerned beause a) I have no idea what the task scheduler is, and b) I normally dual boot in to Linux and Windows 10 (I use Linux for pretty much everything and I would LOVE to didtch windows altogether and just use Linux. I understand that basically everything is designed for maximum performance under Windows, but I have heard some ugly rumors about Microsoft making it harder for computer manufacturers to allow consumers to install Linux / Unix on their machines. This does worry me.

Anyway, thanks for your elaboration.

 

[My PC Hates Me]

A 12600k and Z690 costs roughly the same as a 12700 and B660. There's a performance difference right off the rip, the 12700 beats the 12600k in games and production. The Only benefit of the 12600k is its ability to OC, which is fine for gaming, but introduces the chances of instability, which is Death to production in things like video editing etc, where a single bit-flip can ruin a render, add funky pixels in frames etc. For production, stability is key, and that doesn't happen with OC.

So pay for performance out of the box, or pay for possible speed increases and its inherent hazards. Me personally, I'd not want a glitch happening 3/4 of the way through a 6hr render, forcing a complete start-over.

Words of wisdom!!! Thank you very much.

Yes, the whole thought of overclocking (and under-volting, whatever the heck that is), makes me nervous. And yeah, having a crash when 3/4ths of the way through a long render is not my definition of fun.