FRONTPAGE AMD Ryzen 7 1800X CPU Review

[bassnut]

My take ......

As stated before in one of these threads this is a new platform and there will be bugs.

Looks like, sort term..... in a couple months once the platform matures and stabilizes Intel will have an even better RYZEN to deal with and Long Term if these issues are addressed Intel will have an even better preforming chip to deal with.

Ideally this should have been worked out before release but with all the pressure to release now AMD might have rushed it a bit. I can live with this news.

 

[MNMadman]

My take ......

As stated before in one of these threads this is a new platform and there will be bugs.

Looks like, sort term..... in a couple months once the platform matures and stabilizes Intel will have an even better RYZEN to deal with and Long Term if these issues are addressed Intel will have an even better preforming chip to deal with.

Ideally this should have been worked out before release but with all the pressure to release now AMD might have rushed it a bit. I can live with this news.

Yep. This type of thing should have been a non-issue. But I know things will improve with Ryzen, which is why I'm still excited about the 1800X build I'll be finishing tonight.

 

[Keith Myers]

I understand the comments about AMD needing to release now, maybe too soon. What I don't understand is the thinking behind the decision by AMD management to release now with unfinished OS and BIOS support in place. With the poor industry opinion of their last processors, I would think AMD would have wanted to put out a great product on Day One to diminish those opinions to the dim past. Now with all the printed test reviews of the Ryzen 7 fresh in everyone's mind, they just made it more difficult to overcome the new pre-judged opinions in the current buying public. I doubt very many people will revisit any updated reviews of fixed products when they eventually show up. Which I'm sure they will. First impressions always count. Why did they decide to shoot themselves in the foot again?

 

[Alaric]

Because there weren't any memory bugs with Skylake BIOS. Oh, wait, yes there were. I had a problem with a NEW BIOS not letting me set the XMP profile. Intel has issues with new chips, too. The fact that they haven't made a truly new chip in so long mitigates it somewhat, but it happens to them, too.

 

[wingman99]

Intel dose not support overclocking memory, the motherboard manufactures do the XMP profiles for Bios.

 

[Bluefalcon13]

Either way Alaric's point still stands. So far the biggest problem has been the mobos. It's like the Mobo manufacturers had the b-team on them, then realized a bit late it wasn't cutting it. Look at cdawall's review, problems with the CHVI's BIOS are mentioned in it. Look at all the people NOT buying Ryzen due to Mobo's they want not being in stock anywhere. Almost every review I read mentioned the Mobo's need some tlc in the BIOS department.

 

[EarthDog]

Intel dose not support overclocking memory, the motherboard manufactures do the XMP profiles for Bios.

The IMC is overclocked. Using XMP, the memory is not overclocked. They are rated to run those speeds.

 

[wingman99]

Either way Alaric's point still stands. So far the biggest problem has been the mobos. It's like the Mobo manufacturers had the b-team on them, then realized a bit late it wasn't cutting it. Look at cdawall's review, problems with the CHVI's BIOS are mentioned in it. Look at all the people NOT buying Ryzen due to Mobo's they want not being in stock anywhere. Almost every review I read mentioned the Mobo's need some tlc in the BIOS department.

All Intel and AMD guarantee is running stock. How is AMD doing in that department?

 

[Bluefalcon13]

All Intel and AMD guarantee is running stock. How is AMD doing in that department?

I'm not really sure what you are trying to say?

Does AMD make motherboards now too? Has there been widespread reports of the AMD processors not working due to the processors being faulty?

 

[Alaric]

I was referring to the motherboards/BIOS updates that come with new chips. I don't recall any reviews stating the memory wouldn't reach the motherboard's stated speed.

 

[wingman99]

I'm not really sure what you are trying to say?

Does AMD make motherboards now too? Has there been widespread reports of the AMD processors not working due to the processors being faulty?

What I'm saying is AMD sends the motherboard manufactures stock specification packet for building the motherboard, I was going to sign up with Intel for that, however it was to much paper work. I'm just wondering if Ryzen has any problems running stock that AMD has to update the micro code to fix?

 

[Bluefalcon13]

What I'm saying is AMD sends the motherboard manufactures stock specification packet for building the motherboard, I was going to sign up with Intel for that, however it was to much paper work. I'm just wondering if Ryzen has any problems running stock that AMD has to update the micro code to fix?

Ahhh, it sounded like you were claiming Ryzen was broken or something. As far as I know, the issue is more towards the mobo manufacturers not having the board's uefi fully ready. this goes into it a bit if you havent watched it:

 

[Clutch_Head]

great writeup ATMINSIDE. learned a lot from your very thorough review.

I'll try to answer all questions, I do have AMD contacts who I can ask anything I don't know.

could you ask if they have a Realflow 10 benchmark? preferably one that can run on the demo version?
i'm going to be using Realflow 10 for production work, and i'm considering a Ryzen build.

cheers

 

[IntelNvidia]

Im surprised no one is up in arms over the fact that launch day 1800x is putting up worse numbers than the demo'd 1800x. If your finished product is worse than your engineering samples that is just bad marketing. Means they really did cherry pick a chip JUST for those tests and that they were unrealistic.

As for the mobo and Microsoft support. AMD should have been doing a much better job communicating with the vendors. They were too scared of leak information, which still happened, to allow the vendors good information on their chip.

This war was all about image. They had a terrible image for all the weak cpus they have produced for the last 6-8 years. They wanted to keep things hush hush and release only tid bits of imformation from cherry picked samples to make them look like true contenders. Now the product is out and the dust settles... the support from vendors is bad, the chip performance is worse than promised and worse than the live demos.

I feel this was a bait switch sadly and I will probably never buy another amd product because of it.

 

[Bluefalcon13]

Im surprised no one is up in arms over the fact that launch day 1800x is putting up worse numbers than the demo'd 1800x. If your finished product is worse than your engineering samples that is just bad marketing. Means they really did cherry pick a chip JUST for those tests and that they were unrealistic.

As for the mobo and Microsoft support. AMD should have been doing a much better job communicating with the vendors. They were too scared of leak information, which still happened, to allow the vendors good information on their chip.

This war was all about image. They had a terrible image for all the weak cpus they have produced for the last 6-8 years. They wanted to keep things hush hush and release only tid bits of imformation from cherry picked samples to make them look like true contenders. Now the product is out and the dust settles... the support from vendors is bad, the chip performance is worse than promised and worse than the live demos.

I feel this was a bait switch sadly and I will probably never buy another amd product because of it.

Show us on the doll where the bad AMD man touched you....

But seriously, I don't recall AMD saying anything other than Haswell IPC. Where did they say what you said they said? I'm curious as I have been following Ryzen fairly heavily, and I missed that. Could you please show me?

 

[Alaric]

Show us on the doll where the bad AMD man touched you....

 

[Keith Myers]

Show us on the doll where the bad AMD man touched you....

Show us on the doll where the bad AMD man touched you....

+1 or +1000

 

[wingman99]

Im surprised no one is up in arms over the fact that launch day 1800x is putting up worse numbers than the demo'd 1800x. If your finished product is worse than your engineering samples that is just bad marketing. Means they really did cherry pick a chip JUST for those tests and that they were unrealistic.

As for the mobo and Microsoft support. AMD should have been doing a much better job communicating with the vendors. They were too scared of leak information, which still happened, to allow the vendors good information on their chip.

This war was all about image. They had a terrible image for all the weak cpus they have produced for the last 6-8 years. They wanted to keep things hush hush and release only tid bits of imformation from cherry picked samples to make them look like true contenders. Now the product is out and the dust settles... the support from vendors is bad, the chip performance is worse than promised and worse than the live demos.

I feel this was a bait switch sadly and I will probably never buy another amd product because of it.

AMD mad a good multi-core chip, you can't deny it, only gaming is bad.

 

[ITAngel]

The chip looks amazing but base on this guy review I think I will hold up for music production recording.

 

[Bluefalcon13]

AMD mad a good multi-core chip, you cant deny it, only gaming is bad.

I wouldn't say gaming is bad per say. I would say as a product, Ryzen is great for productivity. It's sufficient at gaming too, but not OMG AMAZING at gaming. Buying it expressly for gaming is not a great idea, but if you use computers for more intensive tasks that can benefit from the additional cores, it does have its advantages.

Case and point:. For an internship I had to write software that computed binary matrixes and stored them to disk. No big deal right? Except each one took a few ms to calculate each. Again, no big deal right? There was only 2^23 matrixes to compute... It took over 24 hrs on an i7 with some minor IO bottlenecking (after a while the disk got its cache saturated). Granted the code I wrote could probably be optimized more than when I wrote it, but it was fairly efficient. I had already done some pretty significant optimizations at that point. We are talking binary matrixes, that when 8 elements (bits) are stored, they are stored as a byte, with very minimal overhead, and consumed 28kilobytes of disk space.

Still would've been faster on 16 threads (and a solid state).