Review Ryzen 7 9700X Reviews

[DAPUNISHER]

MS should bring these gains to Windows 10 as well, it has 1-4 years left of support depending on the type of version/subscription.

People have been switching back to 10 - https://www.digitaltrends.com/computing/windows-11-market-share-april-2024/

It will truly be Hanlon's if they add these opts to 10. It's an actual carrot to get people to use 11.

Oh please, don't play on words.

Is English your first language? Malice is ill will. Money is a financial incentive that can be entirely amoral.

 

[CouncilorIrissa]

Not malice, money.

If the implication is some under-the-table deal between MS and Intel, then no. Have you seen Intel's financials the last few years? They don't exactly have spare cash to throw around.

 

[DAPUNISHER]

If the implication is some under-the-table deal between MS and Intel, then no. Have you seen Intel's financials the last few years? They don't exactly have spare cash to throw around.

Nothing under the table about it. Wintel was not coined without good reason.

EDIT: and you are just another internet rando like the rest of us. Your opinion is no more valid than any of ours.

But since some of you seem intent on obfuscation. Let me point out once again -

Zen on AM5 sees huge gaming uplift in window 11 24H2 🤣

 

[coercitiv]

We're gonna need all new reviews with fresh benchmarks now. Including Z4 vanilla, Z4 3D, and Intel.

Who's gonna give the bad news to all the Steves? Up until now they had to do a fresh review of 9700X every few days or so, now they have to retest everything going back a couple of years. HUB Steve already mentioned lacking the energy to properly retest Intel on the new OS (for now, he did a preview though).

Zen 5 performance is the most elusive I've seen in a long time, the more one tries to measure it, the more errors pop out. The observers effect 101.

 

[Bigos]

Lol, I knew Windows as a platform was FUBAR (which is why I like running Linux whenever possible) but this is a new low.

So for at least 2 years (if not 4) the spectre mitigation vulnerability workarounds were misapplied on AMD CPUs and no one noticed - not MS, not AMD.

I attribute it to incompetence on both sides. Both have screwed up.

 

[Nothingness]

Is English your first language? Malice is ill will. Money is a financial incentive that can be entirely amoral.

I take that as a compliment that you have to ask the question 😉 Though I might have misused the word malice, I was refering to Hanlon's razor: Never attribute to malice that which is adequately explained by stupidity.

That said, I gave a proper example of what could have been seen as a deliberate act to reduce Arm performance, but was only an oversight which was uncovered thanks to people being able to look at the kernel source code. You might rightfully place money above everything, but that doesn't prove MS acted the way you accuse them.

 

[DAPUNISHER]

You might rightfully place money above everything, but that doesn't prove MS acted the way you accuse them.

Agreed. And I don't expect or need any of you to share my opinions. If we all echo chambered that would be immensely boring. And it was indeed a compliment about how well you write in English, I can barely order food or ask where the bathroom is in other languages.

Back to the topic: Looking forward to 3D, Intel, and older Zen testing. So far it looks like a win for AMD and Intel gamers.

 

[DAPUNISHER]

So for at least 2 years (if not 4) the spectre mitigation vulnerability workarounds were misapplied on AMD CPUs and no one noticed - not MS, not AMD.

Is this speculation or fact?

 

[Bigos]

Is this speculation or fact?

This is based on my reading of the situation, so it's still speculation. Let's analyze it a bit.

What if MS knew about that but AMD did not? This would mean that MS withheld the information to affect the AMD CPU performance on purpose, which I believe is what you claim. We can neither prove nor disprove that - just speculate - unless we are provided more first-party information. But even if we assume this to be true, it is still partly AMD's fault for not noticing something that significantly affects the performance of their CPUs for multiple years when using the OS that is the most pervasive in the client segment. And my take is that I find it unlikely MS went so far to gimp the performance of one of its biggest partners on purpose.

If both knew about that it would mean it took so long for AMD to convince MS to fix their stuff. In that case, AMD said nothing about that to the public. It would mean they value the relations with MS higher than the performance of their product compared to the competition. Given how they work around some Windows shortcomings (the special scheduler on X3D and Zen 5) without MS incorporating that in the base system, it sounds strange they would not at least attempt doing something like that in the meantime.

This leaves the last option (not counting the nonsensical one "AMD knew about it but MS did not") that both did not notice. MS might have indeed not cared enough, but AMD should have - as mentioned, we are talking about the OS used by the vast majority of their client market.

Given that this has been fixed already before Zen 5 launch (STX notebooks all shipped with the new OS, apparently), this has probably already been released in the insider Windows builds some time ago. There was not much fanfare about fixing something that affects AMD CPU performance so much from either side. Maybe it was too embarrassing? Maybe the effect was deemed to not be significant enough? Maybe AMD marketing stroke again?

 

[DAPUNISHER]

Maybe AMD marketing stroke again?

On family feud this would be the number one answer. 😀 Certainly near the top of the list. I am going with AMD lacked the market share for MS to effort optimizations for them up until the last couple of years as my top answer. With Wintel shenanigans as my number 2. AMDumb marketing is hard to argue against though. Particularly given the roll they have been on the last couple of months. I would have cleaned house and fired everyone in charge over there already.

 

[ondma]

Not malice, money.

Normally, I agree with most of your posts, but have to disagree here. You could be right, but currently there is no evidence of intentional handicapping of AMD. So I would not just assume this. It is like Intel/MS is guilty until proven innocent.

 

[ondma]

Is there any explanation of why this increase, i.e. the actual mechanism? Is it related to latency improvements?

 

[Hitman928]

Is there any explanation of why this increase, i.e. the actual mechanism? Is it related to latency improvements?

MS kept applying software mitigations for CPU vulnerabilities in AMD CPUs, even in generations where they had hardware mitigations in place (i.e., starting with Zen 3), so the software mitigations were only hurting performance for no added security benefit. It seems that MS turned off the software mitigations for Intel CPUs that had hardware mitigations in place.

 

[DAPUNISHER]

Normally, I agree with most of your posts, but have to disagree here. You could be right, but currently there is no evidence of intentional handicapping of AMD. So I would not just assume this. It is like Intel/MS is guilty until proven innocent.

There is ample precedent for viewing both of those companies with great cynicism. I also appreciate my position on the matter is controversial. I put on the warpaint over the extremely anti consumer behavior Intel has engaged in over Raptor Lake. I ain't taking it off anytime soon. Intel has indeed been shown guilty after refusing to take action and throwing everyone else under the bus, so anything they do or say now is completely untrustworthy. Hence, them paying MS or providing guidance and assistance, or any other method that benefited them and hurt AMD would be 100% on brand.

Microsoft is more a matter of not caring unless you pay them too IMO. If AMD had to provide assistance and work with MS to get things fixed. That means less payroll hours and workload for MS at the very minimum, correct?

People keep referencing Hanlon's but I prefer Occam's. Money is always the simplest explanation when corpos are involved.

MS kept applying software mitigations for CPU vulnerabilities in AMD CPUs, even in generations where they had hardware mitigations in place (i.e., starting with Zen 3), so the software mitigations were only hurting performance for no added security benefit. It seems that MS turned off the software mitigations for Intel CPUs that had hardware mitigations in place.

Thanks for framing it like that. Nothing sus about that at all. /s

I have issued mea culpa numerous times over the years. I will do so again without hesitation. Quite often I have also gotten to subtly or not too subtly celebrate being correct.

The amount of negativity being expressed today after Steve's testing is puzzling. "Big performance was found, get your pitchforks and torches bois!" WTF M8?

 

[Bigos]

If I remember correctly, the HW spectre mitigations have been programmed differently on Intel vs AMD, though that might have been in the previous generations and are unified now. I might be misremembering things, though.

Notice that there is no "one" spectre vulnerability but there are various variants. My current Zen 4 CPU on Linux says:

Vulnerabilities:          
  Gather data sampling:  Not affected
  Itlb multihit:         Not affected
  L1tf:                  Not affected
  Mds:                   Not affected
  Meltdown:              Not affected
  Mmio stale data:       Not affected
  Retbleed:              Not affected
  Spec rstack overflow:  Mitigation; Safe RET
  Spec store bypass:     Mitigation; Speculative Store Bypass disabled via prctl
  Spectre v1:            Mitigation; usercopy/swapgs barriers and __user pointer sanitization
  Spectre v2:            Mitigation; Enhanced / Automatic IBRS; IBPB conditional; STIBP always-on; RSB filling; PBRSB-eIBRS Not affected; BHI Not affected
  Srbds:                 Not affected
  Tsx async abort:       Not affected

The Spectre v2 line is the longest as its mitigation has been the most convoluted to my knowledge.

 

[Josh128]

OK so who else is looking for a deal on Zen 5 today??? 😀😀

 

[ondma]

There is ample precedent for viewing both of those companies with great cynicism. I also appreciate my position on the matter is controversial. I put on the warpaint over the extremely anti consumer behavior Intel has engaged in over Raptor Lake. I ain't taking it off anytime soon. Intel has indeed been shown guilty after refusing to take action and throwing everyone else under the bus, so anything they do or say now is completely untrustworthy. Hence, them paying MS or providing guidance and assistance, or any other method that benefited them and hurt AMD would be 100% on brand.

Microsoft is more a matter of not caring unless you pay them too IMO. If AMD had to provide assistance and work with MS to get things fixed. That means less payroll hours and workload for MS at the very minimum, correct?

People keep referencing Hanlon's but I prefer Occam's. Money is always the simplest explanation when corpos are involved.


Thanks for framing it like that. Nothing sus about that at all. /s

I have issued mea culpa numerous times over the years. I will do so again without hesitation. Quite often I have also gotten to subtly or not too subtly celebrate being correct.

The amount of negativity being expressed today after Steve's testing is puzzling. "Big performance was found, get your pitchforks and torches bois!" WTF M8?

Well, to be fair, he did express skepticism himself and said the rumors appeared "too good to be true".

 

[ondma]

OK so who else is looking for a deal on Zen 5 today??? 😀😀

Not sure I understand your post, if it is sarcasm or what, but the improvement was similar for Zen 4 and Zen 5, so Zen 4 still seems the better value. Compared to Zen 4, Zen 5 still seems a dud for gaming. It will be interesting to see if the new windows build increases application performance, where Zen 5 seems to shine (relatively).

 

[DAPUNISHER]

Well, to be fair, he did express skepticism himself and said the rumors appeared "too good to be true".

Incorrect. He said his data seems too good to be true and is waiting for others to confirm it. I sincerely doubt he goofed this that badly though. If he did it'll be a major L for his reputation as a reviewer. He has the interwebz in quite a stir today.

 

[Golgatha]

Not sure I understand your post, if it is sarcasm or what, but the improvement was similar for Zen 4 and Zen 5, so Zen 4 still seems the better value. Compared to Zen 4, Zen 5 still seems a dud for gaming. It will be interesting to see if the new windows build increases application performance, where Zen 5 seems to shine (relatively).

Maybe think of it like my two Zen 4 systems got a free Zen 5 performance upgrade through software?

 

[IEC]

Maybe think of it like my two Zen 4 systems got a free Zen 5 performance upgrade through software?

Finally, a hype train I can get behind. +10% average free gaming performance without even buying the new chip! Best DLC ever!

 

[KompuKare]

Finally, a hype train I can get behind. +10% average free gaming performance without even buying the new chip! Best DLC ever!

Great but be careful using the term DLC there in case someone tries to monetises this somehow!

 

[GaiaHunter]

Finally, a hype train I can get behind. +10% average free gaming performance without even buying the new chip! Best DLC ever!

So does that mean the 7700/7700X are on par with a 14900K for much lower price, much lower power consumption and no degradation problems, for gaming?

This is before the 10% windows uplift.

 

[DAPUNISHER]

^ It depends on how many games benefit raptor lake as to whether or not vanilla zen catches it for bigger bar better 25 second test loops averages in a select suite of games. If the i5 results are correct, they will certainly beat it in some games now.

I notice Steve suddenly found time to test MS flight sim, very interdasting. Considering if it was part of his normal suite it would benefit Zen 3D more than any other CPU. H estopped because he was whinging about account issues, I see he solved those. Or at least caved to patreon demand.

 

[Hitman928]

So does that mean the 7700/7700X are on par with a 14900K for much lower price, much lower power consumption and no degradation problems, for gaming?
View attachment 106256
This is before the 10% windows uplift.

Still need more testing to validate but potentially, yes. AMD did say in their revised Zen 5 comparison that it was on par in gaming with RPL, so HWUB’s tests at least seem to jive with that being a real possibility.