Intel Core i5-8400 (B360) vs. AMD Ryzen 5 1600 (B350)

[collector1924]

Hi Steve,
Nice article. I have one question for you though.
Are you using a meltdown patched i5 for the benchmark comparisons ?
I am referring to the meltdown patch intel released on Feb 21, 2018.
If you are using the STOCK previously published benchmark numbers from Intel unpatched then this is not a valid benchmark comparison IMO as all Intel CPUS will eventually have this patch applied and it has been shown to increase performance in some areas but greatly decrease it in others especially I/O and some of those benchmarks where Intel Beats AMD may no longer be true.

In any event don't you think that from this point forward that the benchmarks should say if the latest meltdown patches have been installed or not ?

PS:
If you want to see how much performance is affected with the meltdown patch see the conclusions page on this article. In fact some areas show improvement but many show a big performance loss.
Benchmarks that don't use this patch are just not doing fair comparison imo as all INTEL cpus must eventually install it...
"Meltdown & Spectre: Analyzing Performance Impacts on Intel's NUC7i7BNH"
by Ganesh T S on March 23, 2018 4:15 PM EST

Thanks,
Mark

 

[Steve]

Hi Steve,
Nice article. I have one question for you though.
Are you using a meltdown patched i5 for the benchmark comparisons ?
I am referring to the meltdown patch intel released on Feb 21, 2018.
If you are using the STOCK previously published benchmark numbers from Intel unpatched then this is not a valid benchmark comparison IMO as all Intel CPUS will eventually have this patch applied and it has been shown to increase performance in some areas but greatly decrease it in others especially I/O and some of those benchmarks where Intel Beats AMD may no longer be true.

In any event don't you think that from this point forward that the benchmarks should say if the latest meltdown patches have been installed or not ?

PS:
If you want to see how much performance is affected with the meltdown patch see the conclusions page on this article. In fact some areas show improvement but many show a big performance loss.
Benchmarks that don't use this patch are just not doing fair comparison imo as all INTEL cpus must eventually install it...
"Meltdown & Spectre: Analyzing Performance Impacts on Intel's NUC7i7BNH"
by Ganesh T S on March 23, 2018 4:15 PM EST

Thanks,
Mark

All our test systems use the latest versions of Windows with all the updates applied as well as the latest drivers and firmware updates. I'm not sure where you're pointing me for information on the impact Meltdown & Spectre had on performance, but I covered this topic in quite some depth, so I'm aware.

 

[Stark]

Good read Steve, thanks for the hard work.
Ignore the trolls for each one pointing finger I believe there are ten who know about your integrity.
Just have fun. Cause so much hardware you got to play with man.

 

[Nightfire]

Steve, before digging into the substance, I must correct you on the number of Amazon reviews. The 400 reviews you see for the 8400 is for all CFL combined, ie the 8600k and 8700k have the exact same number.

The 400 reviews you see for the r5 1600 is for the r5 1600 ONLY meaning Ryzen is really doing well there.

 

[Steve]

Steve, before digging into the substance, I must correct you on the number of Amazon reviews. The 400 reviews you see for the 8400 is for all CFL combined, ie the 8600k and 8700k have the exact same number.

The 400 reviews you see for the r5 1600 is for the r5 1600 ONLY meaning Ryzen is really doing well there.

I t w a s a j o k e

Good read Steve, thanks for the hard work.
Ignore the trolls for each one pointing finger I believe there are ten who know about your integrity.
Just have fun. Cause so much hardware you got to play with man.

Thanks mate, appreciate that and I do enjoy playing around with the hardware

 

[Nightfire]

Great review Steve!

I still think the 8400 is a good cpu but it sort of got a free pass on initial review with the z370.

"B360 will be out soon and you will get the same performance for $50 less"

Well, B360 took a LONG time to release, took away MCE, and nerfed the memory support. In the end, it actually made the R5 look even more appealing.

Furthermore, the R5+ is right around the corner..

 

[shakattack]

So basically, AMD is still king for productivity and Intel is still king for gaming - provided you have a GPU fast enough to take advantage of it.

Older DX11 titles maybe. All newer games utilize the extra cores for physics and AI and are better threaded. Besides, AMD offers an upgrade path to 2700X (air-cooled 4.3GHz all cores) and even the 7nm Zen2 next year. Cant say the same for Intel.

 

[CheekyP]

Great review Steve!

I still think the 8400 is a good cpu but it sort of got a free pass on initial review with the z370.

"B360 will be out soon and you will get the same performance for $50 less"

Well, B360 took a LONG time to release, took away MCE, and nerfed the memory support. In the end, it actually made the R5 look even more appealing.

Furthermore, the R5+ is right around the corner..

MCE doesn't work on non K CPUs so it's not relevant.

It was known all along that DDR4 2666 would be the limit on non Z370 mobos.

I don't see how that makes Ryzen any more appealing? The value aspect of the 8400 improved with B360, especially if you are a gamer wanting to get the highest possible framerates for the least amount of money.

The fact that an i5 8400 / DDR4 2666 combo outperforms an overclocked 4GHz R5 1600 / DDR4 3200 for gaming while drawing half the power is not lost on me.

You save about $50 on the platform with Intel (due to cheaper memory, plus the cost of a decent HSF to overclock a 1600 to 4GHz)

That $50 can go towards a GPU upgrade, you can go from a 1070 to a 1070 Ti for example with the money you saved.

 

[CheekyP]

Older DX11 titles maybe. All newer games utilize the extra cores for physics and AI and are better threaded. Besides, AMD offers an upgrade path to 2700X (air-cooled 4.3GHz all cores) and even the 7nm Zen2 next year. Cant say the same for Intel.

You mean newer games like Far Cry 5 or Assasins Creed: Origins?

I don't know how well FC5 scales with threads but I know for a fact that AC:O scales well with additional cores/threads and guess what? Ryzen is still way behind.

An upgrade path to a 2700X sounds appealing, until you realise that it can't even outperform a 8400 in gaming, let alone a 8700K.

AMD has a LONG WAY to go to beat Intel at gaming. They are ahead for productivity, sure, at least when it comes to Ryzen 5 vs Core i5.

 

[NightAntilli]

The fact that an i5 8400 / DDR4 2666 combo outperforms an overclocked 4GHz R5 1600 / DDR4 3200 for gaming while drawing half the power is not lost on me.

You cannot conclude that. The power consumption test was not done under gaming. And the main reason that the R5 1600 loses under gaming is because the majority of its threads are not used, meaning its power consumption under gaming will be significantly lower than under Blender.

@Steve... It might be an idea to test power consumption specifically under gaming and add that to the article, or at least for future articles. People are drawing conclusions here based on hasty generalization...

 

[Daithi]

I don't know how well FC5 scales with threads

It's pretty well multithreaded, however Dunia is derived from Cryengine and they both have the same issue, they have a main thread and a render thread and they absolutely hammer one or the other and it bottlenecks everything else. A kingdom come deliverance dev explained it better on Digital Foundry's PC analysis video.

 

[CheekyP]

You cannot conclude that. The power consumption test was not done under gaming. And the main reason that the R5 1600 loses under gaming is because the majority of its threads are not used, meaning its power consumption under gaming will be significantly lower than under Blender.

@Steve... It might be an idea to test power consumption specifically under gaming and add that to the article, or at least for future articles. People are drawing conclusions here based on hasty generalization...

Fair points, I would like to see gaming power consumption numbers too, though I'm sure the 8400 would still be significantly lower, though perhaps not to the extent shown in the Blender tests.

 

[Nightfire]

I t w a s a j o k e

Was it Steve? Was It? Come on now...

 

[CheekyP]

It's pretty well multithreaded, however Dunia is derived from Cryengine and they both have the same issue, they have a main thread and a render thread and they absolutely hammer one or the other and it bottlenecks everything else. A kingdom come deliverance dev explained it better on Digital Foundry's PC analysis video.

Just did a little searching on this topic and came across this:
http://www.guru3d.com/articles-pages/far-cry-5-pc-graphics-performance-benchmark-review,8.html

It appears that 6C/6T is the 'point of diminishing returns' when it comes to Far Cry 5.

I think most games still have that problem of a dominant thread limiting performance, don't think that will go away anytime soon as games are a lot harder to code for parallelism than say, rendering or encoding tasks.

 

[Nightfire]

Great review Steve!

I still think the 8400 is a good cpu but it sort of got a free pass on initial review with the z370.

"B360 will be out soon and you will get the same performance for $50 less"

Well, B360 took a LONG time to release, took away MCE, and nerfed the memory support. In the end, it actually made the R5 look even more appealing.

Furthermore, the R5+ is right around the corner..

MCE doesn't work on non K CPUs so it's not relevant.

It was known all along that DDR4 2666 would be the limit on non Z370 mobos.

I don't see how that makes Ryzen any more appealing? The value aspect of the 8400 improved with B360, especially if you are a gamer wanting to get the highest possible framerates for the least amount of money.

The fact that an i5 8400 / DDR4 2666 combo outperforms an overclocked 4GHz R5 1600 / DDR4 3200 for gaming while drawing half the power is not lost on me.

You save about $50 on the platform with Intel (due to cheaper memory, plus the cost of a decent HSF to overclock a 1600 to 4GHz)

That $50 can go towards a GPU upgrade, you can go from a 1070 to a 1070 Ti for example with the money you saved.

Gotta love that Cheeky Math: 117w is half of 155w. Even then, the R5 was slightly faster on blender stock.

3200 mhz ram is $50 more than 2666 mhz ram huh? Where do you live, North Korea??

 

[CheekyP]

Gotta love that Cheeky Math: 117w is half of 155w. Even then, the R5 was slightly faster on blender stock.

3200 mhz ram is $50 more than 2666 mhz ram huh? Where do you live, North Korea??

I was comparing the overclocked 4GHz 1600 to the stock 8400, I clearly said that in my post, perhaps reading comprehension is your problem, not my 'cheeky maths'.

I also said the cost of RAM *plus* the need for an aftermarket cooler to hit 4GHz on a 1600. Again, your reading comprehension fails you. 3200 is about $25 - $30 more expensive than 2666, and I factored in a $30 cooler like the popular EVO 212. Of course, you can forgo the aftermarket HSF and stick with the Wraith Stealh, which isn't a bad stock HSF at all, but you'll most likely be limited to around 3.8GHz with that HSF.

My point remains - for gaming, a Ryzen 1600 + DDR4 3200 is slower, more expensive, and less power efficient than an i5 8400 + DDR4 2666.

 

[NightAntilli]

Fair points, I would like to see gaming power consumption numbers too, though I'm sure the 8400 would still be significantly lower, though perhaps not to the extent shown in the Blender tests.

Some other sites give some insight... Although I'm not sure how they calculated it...
http://www.tomshardware.com/reviews/intel-coffee-lake-core-i5-8400-cpu,5281-10.html

 

[Nightfire]

I was comparing the overclocked 4GHz 1600 to the stock 8400, I clearly said that in my post, perhaps reading comprehension is your problem, not my 'cheeky maths'.

I also said the cost of RAM *plus* the need for an aftermarket cooler to hit 4GHz on a 1600. Again, your reading comprehension fails you. 3200 is about $25 - $30 more expensive than 2666, and I factored in a $30 cooler like the popular EVO 212. Of course, you can forgo the aftermarket HSF and stick with the Wraith Stealh, which isn't a bad stock HSF at all, but you'll most likely be limited to around 3.8GHz with that HSF.

My point remains - for gaming, a Ryzen 1600 + DDR4 3200 is slower, more expensive, and less power efficient than an i5 8400 + DDR4 2666.

Well then it is your Cheeky Logic. You assume that the overclocked 1600x will use that same wattage in gaming. Clearly, the r5 is not at 100% cpu utilization during those games, especially overclocked, so yeah Cheeky Logic.

Maybe you can hit 3.8ghz on a stock maybe 4.0ghz - whatever. Anyhow, you ignore all tests except for the 1080p gaming, which seems to carry more weight than all aspects of computing combined. Cheeky Logic!

 

[Capaill]

For pure gaming, Intel wins. Even more so if you can budget up for the unlocked CPUs where clock speed is a huge benefit for gaming.
For pure productivity, AMD wins. 12 threads are better than 6.
For a mix of productivity and gaming (especially game streaming), I think AMD wins.
But for budget / value gaming, where you won't have a GTX 1080 in the box, then it's a bit of a crapshoot. Right now, probably go for Intel but try to wait and see what the Ryzen2 chips offer.
Personally I am mostly gaming but still went for the Ryzen 1600 (OC to 3.8) as I don't game above 60fps and will probably upgrade to a future Ryzen chip on the same motherboard.

 

[Nightfire]

Some other sites give some insight... Although I'm not sure how they calculated it...
http://www.tomshardware.com/reviews/intel-coffee-lake-core-i5-8400-cpu,5281-10.html

Ahhh, don't show Cheeky this. It makes way too much sense. How dare you compare power consumption during gaming while making a comparison of power consumption during gaming.

 

[CheekyP]

Well then it is your Cheeky Logic. You assume that the overclocked 1600x will use that same wattage in gaming. Clearly, the r5 is not at 100% cpu utilization during those games, especially overclocked, so yeah Cheeky Logic.

Maybe you can hit 3.8ghz on a stock maybe 4.0ghz - whatever. Anyhow, you ignore all tests except for the 1080p gaming, which seems to carry more weight than all aspects of computing combined. Cheeky Logic!

Your fanboy logic seems to win out over everything else

I'm not disputing AMD's supremacy in productivity, Ryzen's value here is unparalleled and if my workload was a mix of work and gaming I would definitely go with Ryzen 1600 over an 8400.

But for a gamer, 8400 wins, hands down. Its cheaper and faster, and more efficient.

 

[Strawman]

You mean newer games like Far Cry 5 or Assasins Creed: Origins?

I don't know how well FC5 scales with threads but I know for a fact that AC:O scales well with additional cores/threads and guess what? Ryzen is still way behind.

An upgrade path to a 2700X sounds appealing, until you realise that it can't even outperform a 8400 in gaming, let alone a 8700K.
AMD has a LONG WAY to go to beat Intel at gaming. They are ahead for productivity, sure, at least when it comes to Ryzen 5 vs Core i5.

Actually, AC origins may have high CPU usage on multicores but it doesn't do anything with the extra cores / threads. My 8700k performs almost identical to a 7700k on this game. We are talking about a 3 to 5% difference. That's not what I'd call great scaling.

 

[CheekyP]

Ahhh, don't show Cheeky this. It makes way too much sense. How dare you compare power consumption during gaming while making a comparison of power consumption during gaming.

https://www.gamersnexus.net/hwreviews/3086-intel-i5-8400-cpu-review-2666mhz-vs-3200mhz-gaming/page-2

40 watts 8400 vs 85 watts for the overclocked 1600

 

[Strawman]

But for a gamer, 8400 wins, hands down. Its cheaper and faster, and more efficient.

No, it's definitely not more efficient. Woah..where did that come from?

PS1. It came from Firestrike? Is it a game you play a lot? I'll have to try it.

PS2. Efficiency is measured by performance / watt, not by watt alone. You have no idea do you? In Firestrike Ryzen consumes more power but it also gets 50% better performance.

 

[CheekyP]

No, it's definitely not more efficient. Woah..where did that come from?

PS1. It came from Firestrike? Is it a game you play a lot? I'll have to try it.

Why don't we get Steve to test gaming power consumption? I'm sure the 8400 will be lower, so yes, more efficient.

Please explain to me how a 1600 that draws more power and delivers lower frame rates is the more efficient chip for gaming?