News Which PC Boots Up and Shuts Down Faster: AMD or Intel?

[Admin]

We put Intel's Z490 platform head-to-head with AMD's X570 to see which system boots the fastest.

Which PC Boots Up and Shuts Down Faster: AMD or Intel? : Read more

 

[563REt]

We put Intel's Z490 platform head-to-head with AMD's X570 to see which system boots the fastest.

Which PC Boots Up and Shuts Down Faster: AMD or Intel? : Read more

The CPU does not make any difference, its the speed of your storage device. If your using a hardrive then your windows will boot up slower and be slower in general. Use an SSD or an M.2 NVme SSD for a faster Windows Experience.

 

[Aaron44126]

The CPU does not make any difference, its the speed of your storage device. If your using a hardrive then your windows will boot up slower and be slower in general. Use an SSD or an M.2 NVme SSD for a faster Windows Experience.

I get where you're coming from, but for the purposes of what was being measured in this article... They used the same storage device in all of the tests so clearly that is not the only factor.

 

[fevanson]

It looks like AGESA MOBO firmware from AMD's side takes a bit longer to initialize than Intel, otherwise I believe the actual boot up times will be almost 1:1.

 

[AndrewJacksonZA]

Slow news day?

 

[madmatt30]

Probably the most pointless test ever in all fairness.

If you can't handle waiting 15 seconds to get into windows in a useable state there's something inherently wrong.

My xbox one takes longer to boot than that!

 

[Kamen Rider Blade]

Considering I'm used to over 1 minute boot times from older systems of yester year.

These speeds are largely meaningless.

And I rarely reboot, shut down, or cold boot.

Most of the times my system is just on.

 

[cmoswizard]

This analysis ignores the many times that Microsoft decides to borrow your computer for updates both during boot up and shutdown. Average those in for an infrequent user and Windows loses hands down to Linux.
AMD vs. Intel? Probably the same as stated here.

 

[jgraham11]

Wow, pretty crappy review. Another example of Tom's favouring Intel

Why didn't you use an M.2 drive, they are 4-8 times faster than the SATA drive they chose. Notice the article never says SATA drive, I just searched it and the only reference is "Disabled the third-party SATA controller ".

Could it have been because Intel doesn't support PCIE 4.0, which would have shown Intel being slower.

Also, price wise, which is how all comparisons should be done: the 10600k would be price equivalent to the 3700X not the 3600X.

 

[NightHawkRMX]

They didn't use a SATA drive. They used an intel Optane PCIe SSD. Very fast though id assume PCIe 3 limited.

I have built several AMD and intel machines and I can say, AMD usually does take a little longer to boot. But honestly it's not a metric that matters. 5-10 seconds more? Who cares.

 

[PCWarrior]

So Intel boots 50% faster than AMD in Windows and shuts down 72.7% faster in Linux. AMD fanboys feeling butthurt when seeing AMD losing badly, never fails to make me laugh.

Wow, pretty crappy review. Another example of Tom's favouring Intel

Why didn't you use an M.2 drive, they are 4-8 times faster than the SATA drive they chose. Notice the article never says SATA drive, I just searched it and the only reference is "Disabled the third-party SATA controller ".

They didn’t use a SATA drive. They used the PCIE Intel Optane Drive here. That’s pretty much the fastest drive money can currently buy when it comes to latency and responsiveness.

 

[madmatt30]

So Intel boots 50% faster than AMD in Windows and shuts down 72.7% faster in Linux. AMD fanboys feeling butthurt when seeing AMD losing badly, never fails to make me laugh.


They didn’t use a SATA drive. They used the PCIE Intel Optane Drive here. That’s pretty much the fastest drive money can currently buy when it comes to latency and responsiveness.

AMD fan boys butthurt, are you living in the past or something?

No one actually cares about a 6 second difference in boot time, it's totally irrelevant.
Its the bios initialisation that takes longer on ryzen based boards, nothing at all to do with the cpu.

 

[seanwebster]

Considering I'm used to over 1 minute boot times from older systems of yester year.
These speeds are largely meaningless.
And I rarely reboot, shut down, or cold boot.
Most of the times my system is just on.

Leaving it on when you're away could potentially be a security risk. I know this first hand as my system was compromised a few years ago in the middle of the night. Fortunately, I was on it at the time and watched as the person attempted some bank transfers with failure.

Also, faster boot times are very much noticeable to me. I'm not an average user, but in my experience, I enjoy having my system respond to my requests as fast as possible.

This analysis ignores the many times that Microsoft decides to borrow your computer for updates both during boot up and shutdown. Average those in for an infrequent user and Windows loses hands down to Linux.
AMD vs. Intel? Probably the same as stated here.

Both systems were tested with the latest OS Images. Updates on both OSes were installed before testing began. Both updated to the latest UEFI.

Wow, pretty crappy review. Another example of Tom's favouring Intel
Why didn't you use an M.2 drive, they are 4-8 times faster than the SATA drive they chose. Notice the article never says SATA drive, I just searched it and the only reference is "Disabled the third-party SATA controller ".
Could it have been because Intel doesn't support PCIE 4.0, which would have shown Intel being slower.
Also, price wise, which is how all comparisons should be done: the 10600k would be price equivalent to the 3700X not the 3600X.

I'm just a storage editor reporting on some findings between the two test systems I have here. It's not a review.

I used an Intel Optane SSD 905P because it is the fastest booting SSD I have. The current-gen Phison E16-based PCIe 4.0 x 4 SSDs with 3D NAND flash boot slower than Intel's 3D XPoint on both platforms.

After using AMD's X570 chipset and an R5 3600X for reviews for over a year now I gotta say, I do much prefer Intel's platforms for my use. I have used Intel-based platforms for reviewing SSDs for almost a decade. I've experienced my fair share of storage-related issues on AMD's platforms (both past and current) that have been a pain to deal with in my workflow at times. I can only hope there are fewer on the next X670 platform or if things go well, I can transition back into an Intel system for Gen4/5 testing. Plus, faster boot and restart times mean less time waiting on a machine. I can swap hardware faster and since I reboot often for benchmarking, the seconds add up.

CPUs were chosen since both are 6C/12T.

 

[JamesSneed]

So what we have seen with this is that boot times are basically meaningless due to how fast they are now. AMD lost by 4.5 seconds which in the grand scheme of things is noise. I have a feeling its more about the BIOS than CPU's, drives, etc.

 

[King_V]

I'll admit that faster boot times are nice to have, and I was surprised when I casually noticed that my Haswell era Dell with 2.5" SATA SSD seems to boot faster than my son's Ryzen 1600AF with NVMe.

That said, I don't think it would ever affect my choice as to whether to go with AMD vs Intel as a platform, versus, say, price-performance ratio while I'm using the system.

Still, it would be nice if AMD could manage to get board initialization faster, as long as there were no downsides to it.

 

[madmatt30]

I'll admit that faster boot times are nice to have, and I was surprised when I casually noticed that my Haswell era Dell with 2.5" SATA SSD seems to boot faster than my son's Ryzen 1600AF with NVMe.

That said, I don't think it would ever affect my choice as to whether to go with AMD vs Intel as a platform, versus, say, price-performance ratio while I'm using the system.

Still, it would be nice if AMD could manage to get board initialization faster, as long as there were no downsides to it.

If you go to the trouble of installing windows in UEFI mode and enabling ultra fast boot in asrock boards it can halve bios initialisation time.
That seems to be the main issue on my rig at least.

Mine takes around 10 seconds to get past bios to windows loading screen and then only a further 10 seconds to boot into windows, load all background processes and boot steam big picture mode.

20 seconds I can live with to be honest.


Hell, sometimes I boot the xbox one up after a couple of weeks and it wants to do a 5gb update!!

I'm with you though in the old systems, I have both an intel lenovo Ivybridge and an fx8350 rig that either beat or match the ryzen on boot time on a standard ssd when the ryzen rig has an sx8200 nvme drive in there.

 

[hertzsae]

Leaving it on when you're away could potentially be a security risk. I know this first hand as my system was compromised a few years ago in the middle of the night. Fortunately, I was on it at the time and watched as the person attempted some bank transfers with failure.

Also, faster boot times are very much noticeable to me. I'm not an average user, but in my experience, I enjoy having my system respond to my requests as fast as possible.


Both systems were tested with the latest OS Images. Updates on both OSes were installed before testing began. Both updated to the latest UEFI.


I'm just a storage editor reporting on some findings between the two test system's I have here.

I used an Intel Optane SSD 905P because it is the fastest booting SSD I have. The current-gen Phison E16-based PCIe 4.0 x 4 SSDs with 3D NAND flash boot slower than Intel's 3D XPoint on both platforms.

After using AMD's X570 chipset and an R5 3600X for reviews for over a year now I gotta say, I do much prefer Intel's platforms for my use. I have used Intel-based platforms for reviewing SSDs for almost a decade. I've experienced my fair share of storage-related issues on AMD's platforms (both past and current) that have been a pain to deal with in my workflow at times. I can only hope there are fewer on the next X670 platform or if things go well, I can transition back into an Intel system for Gen4/5 testing. Plus, faster boot and restart times mean less time waiting on a machine. I can swap hardware faster and since I reboot often for benchmarking, it the seconds add up.

CPUs were chosen since both are 6C/12T.

When I first read the article, my first thought was "I wonder how much Intel paid to sponsor this post". Then I thought that would be too blatant, so I thought "I bet some marketing people at Intel looking for good news subtly suggested this as an interesting topic to an unsuspecting journalist as easy fodder for an article without mentioning that it just may happen to turn out in Intel's favor".

Then I read your description and it all made sense. I remember some platforms that I used to work do dev work on. I worked in a software area that wasn't usually CPU intensive and didn't care what hardware it was run on. Each recompile needed a reboot to test. The box we sold was updated from a BIOS based system to a much faster UEFI based one. The BIOS system booted in a couple minutes while the UEFI one took fifteen. Customers rebooted less than once a year, so it didn't matter to them, they just needed speed. But my team and I hated the new box.

With your job and everyone praising AMD (for good reason) and knowing that AMD is worse for you, the article really makes sense. If your job involves testing hardware and booting machines all day long, you don't care about cores and frame rates. You care about boot time. This article is good for the few of us that also care about boot time (thankfully not me at the moment).

 

[hotaru.hino]

The CPU does not make any difference, its the speed of your storage device. If your using a hardrive then your windows will boot up slower and be slower in general. Use an SSD or an M.2 NVme SSD for a faster Windows Experience.

CPU speed does matter, but not in this case. Think about it, booting a system doesn't involve just moving data from storage to RAM, you still have to initialize all of the kernel, services, and any additional hardware based on what was detected. A faster CPU can help in this regard. Program loading requires a minimum amount of CPU work before the user can do anything with it.

The reason it doesn't matter in this case is the performance difference between CPUs has to be significant enough. I recall running a test for craps and laughs where I paired an SSD with a crippled processor (basically an i5 with half the cores and and speed) and an HDD to a faster processor. At least on a clean setup, the faster processor managed to boot in the same amount of time as the slower one with the SSD.

 

[eichwana]

It would be interesting to see if this is the CPU affecting this, or the motherboard. Is this a case of the BIOS initialising things, is it a case of the BIOS itself just being slow?

I'm reminded of an 1156 Gigabyte board back about 10 years ago that had a "vigorous" POST. It would stay on a black screen for about 30 seconds, clicking away, before the BIOS image even showed up. It was RMA'd multiple times (I worked in a computer supplier), but Gigabyte maintained that it was by design.

 

[DZIrl]

I really do not care about boot. Even if times are in minutes who cares! PC is booted once. Eg I reboot my PC every Monday only cause of Windows. Now during Covid I'm rebooting once in a month.

Booting speed mainly depends on BIOS and various options. Eg Gigabyte x570 has several boot modes but none except slow one is suggested.

 

[hotaru.hino]

On a side note, it's amusing to see comments like "what's the point of this article" or "this article doesn't apply to me"

I find articles about mundane things we don't really think about interesting. This goes in line with things like reviewing lower-end parts or whatnot. I may not use them or have a care for them, but someone else might and having that data could help me point them to a better direction.

Besides, articles about the latest and greatest high-end tech are a dime a dozen. Sure I'm excited about the new Ampere cards and what AMD is going to do, but articles about them are expected. Expected things get boring after a while.

 

[Flayed]

Anything under 30s is fine by me. If my job involved swapping out hardware and restarting a lot then I could see how it would be nice if it was faster.

 

[os2wiz]

It looks like AGESA MOBO firmware from AMD's side takes a bit longer to initialize than Intel, otherwise I believe the actual boot up times will be almost 1:1.

With the soon to be released Zen 3 chips memory latencies will be reduced considerably, so expect much of the difference to evaporate shortly. Zen 3 has a much improved memory controller and Infinity Fabric speeds are also much improved within the cpu.

 

[hertzsae]

With the soon to be released Zen 3 chips memory latencies will be reduced considerably, so expect much of the difference to evaporate shortly. Zen 3 has a much improved memory controller and Infinity Fabric speeds are also much improved within the cpu.

I doubt that AMD's longer boot times have to do with hardware slowness. When SSDs came out, boot times didn't initially improve the way people expected (they improved, just not as much). I seem to remember Intel and MS putting a lot of effort into boot times around the time the 'Ultrabook' branding came to be. That brought boot times close to what people expected with SSD drives in their notebooks. The speed came from improving the way that hardware was initialized. I speculate that Intel has simply spent far more effort optimizing the firmware initialization routines for their chipsets and related hardware.

As seen in the comments here, most enthusiasts don't even care that much about boot times. Intel has the money to throw at problems like this that will improve the experience for 1% of their user base. AMD doesn't have those kind of resources. That's also why Intel can make such good network chips and according to a comment made by this author, their storage controllers seem to be less problematic as well. AMD's budget means they have to put the bulk of their resources on what matters to the 99%. Despite their current misteps, Intel can match those resources and still have plenty of engineering resources left to do things like optimize firmware initialization and make bomb proof network and storage controllers.

 

[Integr8d]

Slow news day?

I can't believe this article made it past the editor's desk.