kksnowbear

Your assumptions about how I built the systems? Wrong. Your initial assertion that you can't use four 6000MT/s RAM modules at a CAS less than 40? Also wrong. Your 'bet' regarding custom memory timings etc? You'd lose. If I can provide a better deal on performance that is indistinguishable using components that at the time were *far* less expensive... ...well, again, I don't imagine I'd be doing builds (and second and even *third* builds) for people who keep on coming back, unless I'm providing a service they certainly seem to value. Interesting. Hey I know... a compromise: You keep on telling folks it can't be done. Meantime, I'll keep on doing it. PS Just checked; the local MicroCenter has several AM5 bundles on sale right now... ...and exactly none of them include CAS30 modules. Regardless of which CPU, and even though all are 6000MT/s modules, none are CAS30. That's probably one of the biggest brick and mortar outfits the US. TBH I don't think the 99% figure you cited earlier is anywhere near accurate. You have sales figures or some other reference?

No, I'm afraid you're not close enough. (Yes the B670E-F was a typo, now corrected, thanks). I'm not using custom timings or any such thing. And IIRC the memory (at least to the best of my recollection, since three of the four systems are not here anymore for me to confirm) uses Samsung devices. I'm not recommending anything special to these people, and I'm not doing it to their systems either (unless they explicitly specify such things, but that doesn't apply to these AM5 builds.) No custom values, no adjusted voltages, no modified timing values...nope. Just SPD values, set up properly via standard BIOS profile. So you'd lose that bet. And I'll say it one more time: You didn't say anything about CAS levels until later in the thread - at which point you tried to invoke the "ideal" 6000MT/s CAS30 - which again, as I already said, the benefits of which are arguable at best. I would go so far as to say no one would ever be able to *see* the difference in two systems at CAS30 and CAS32 or 36 RAM, assuming everything else was identical in a proper blind test. I'd bet money on it, in fact. And now you have latched on to CAS30 and won't let it go. Who said it had to be CAS30 (other than you)? I didn't say that - and neither did you, initially...in fact, you said: "I'll rephrase then. It will maybe run at AMD EXPO if that's below 5600Mhz CL40 speed/latency" I can assure you (and did) all 4 of the builds I'm referring to are running 6000MT/s and less than CAS40. You were not involved in these builds, and thus cannot speak to other factors that *definitely* played a part (such as cost, for example; maybe that's why the OP's builder advises as he does - can't speak for him though). In my case, the clients are all perfectly happy; the machines run fine, from initial testing through today. All have been in service for months now, and one is over a year ago (March '23). I deliver what I represent, including everything I'm citing here, and for less cost than any of the clients are likely to find on their own. And no, to be accurate, I do not do it for the money - I'd starve to death if it was my livelihood. I doubt these people would work with me if they didn't agree that I deliver. Why does it bother you? (Other than the fact that it shows your comment about four memory modules wasn't accurate, I mean - which it most assuredly does).

Your advice (concerning four modules) was inaccurate. I offered a comment on first hand experience that showed this. Simple. Your assumptions concerning CAS30 were just that, and I've seen no proof in this thread (or really anywhere) about what "99%" of people are using. Sounds like opinion itself, not proven fact. (I actually doubt 99% is accurate) Everything after that was your personal attacks, and my responding to the posts you made. You can't hide behind accusations of "derailing" and "hijacking" every time just because you're being proved wrong. This is really just an effort to make it seem like I'm somehow breaking forum rules when all I'm doing is discussion. Sorry I'm not obligated to agree with you, especially when your comments are misleading. It's not "derailing a thread"; it's called 'discussion'. Just because you dont agree doesn't mean I'm wrong, nor that there's anything wrong with my comments.

Personal insults have no place in discussion forums. Yet when all else fails you, that's your "go to". Again my clients are happy. And the number is increasing. (Yup, including those right here on this forum). Must be doing something right. Unless you're saying they're all stupid, that is.

Anyone who's worked with these things understands - or should - that by lowering CAS levels enough with high enough speeds, you can make RAM unstable. (By the way this isn't new, and certainly isn't limited to AM5). You started off by saying you can't use 4 modules, which is not correct. I've done it, several times, with no problems. Later in the discussion, you added mention of CAS levels, which (of course) if you try to run higher speeds with lower CAS levels, eventually you'll have problems. Again, not limited to AM5 or that IMC alone. Eventually, you settled on 6000 CAS30 as 'ideal' which by your own admission you can't even make work with four modules. How's that ideal? "Ideal" for me means it actually works. The ones I built absolutely do work as rated and designed. Not much point in carrying on about an ideal that is practically unattainable. (By the way, AMD has published guidance on this subject long ago, and it's all entirely within design specs). Trying to run speeds too high with CAS too low is bound to cause problems - not news, and certainly not limited to AM5. I selected components that would deliver the result to the client - which is also what I said earlier: Who does the work is also a factor. If the OP is paying someone, then it's up to that person to deliver or make it right. My clients are happy.

You didn't mention CAS until later. The nature of 'edits' means the posts aren't the same any more, and I'm not chasing after that, thanks. But the fact is you didn't mention CAS until later, and you've also progressively lowered the value to a point (30) where yes, it would likely be problematic. (You also introduced the term 'ideal', the perceivable benefit of which is arguable). But I haven't done that, I haven't promised anything of the sort to a client - because I know better. Let me make it simpler: Claiming you can't use four modules - wrong. (That's what I called you on in the first place) Saying you can't use four modules faster than 5600 under CAS40 - still wrong. Done it on all four builds, 100% success. Finally saying 64G at 6000 CAS30...yeah, now you've finally whittled it down to something that is a bad idea. I wouldn't suggest that - didn't say I would, didn't say I have. See, what you've done is progressively made the specs you're stating more likely to fail, by changing the technical factors.

Nobody said you can't edit posts. The problem is changing your technical factors as the discussion progresses. This doesn't mention CAS levels at all (in the paragraph where you were discussing four modules). Later you mentioned CAS below 40, and finally (now) you're very adamantly stuck on CAS30. Yes, if you pursue some things that are just bad ideas, you're creating a situation that is ill-advised. Applies to everything. I don't do that, hence my 100% success - no "headaches" at all. Not for me, not for my clients. PS: My results are 100% accurate as stated. You don't need to start with the insults again. I knew the result I'd be getting at the onset, that's what I represented to the clients - and that's what I delivered. 100% of the time.

The edits I'm referring to were not simple typos. They altered the technical values being discussed, essentially as i said already: First it was "You can't run four modules", then we edited it to "Well, you can but they won't run at anything faster than 5600 CAS40..." And finally "Well it'll run 5600 but not at the ideal 6000MT/s CAS30". (I'm paraphrasing, but these are the accurate nature of the edits in general). As I mentioned, these edits are continuously changing the rules mid-game to a standard that is progressively less attainable (regarding speed and CAS). By the last edit, you're describing something that I know isn't going to work, and therefore I'd never sell it to a client. This is also fairly simple: Client wants 64G, I put it in their build - within the technology's limits, of course, regarding speeds. But, as we all know, having too little RAM is far (far) worse than having speeds that are slightly less. BIOS versions are generally updated to whatever is current during the build. I'm sure it's varied over time, and I'm equally sure that some of my clients are capable of updating BIOS themselves, while most of them won't do it.

Several members of this forum have done exactly that (trusted me to do their builds) and they're all happy, thanks. It's not magic. It's understanding the technology and working with what it is. Simple. PS: Pretty sure translators aren't making you go back and change your posts. The translator may not always be accurate, but the edits of themselves were not initiated by a translator.

All four builds I'm referring to are running 6000MT/s modules at the rated speed and CAS latency less than 40. These were all Asus boards. Three are B650E-F models, my own desktop is a X670E-F ROG Strix Gaming model. This is kinda what I meant about choosing carefully the components and builder. I had no problem getting these to work as designed. Invoking factors such as what's "ideal" obviously doesn't make a lot of sense when doing so results in a build that doesn't work. See, these edits are continuously changing the rules mid-game to a standard that is progressively less attainable. First it was "You can't run four modules", then we edited it to "Well, you can but they won't run at anything faster than 5600 CAS40..." And finally "Well it'll run 5600 but not at the ideal 6000MT/s CAS30". (I'm paraphrasing, but these are the accurate nature of the edits in general). That's the problem: Yes, if the builder insists on specs that are simply not attainable, then a failure is pretty much assured from the beginning. A good builder understands the limits of the technology - for better or worse, whether right or wrong - and works within those limits to achieve a high success rate. So far, four of four (100%) that I've done have worked exactly as planned, exactly as rated, and for clients who are 100% pleased with the outcome. Yes, we'd all prefer to have hardware that runs faster, better etc...but as the old saying goes "Wish in one hand, sh*t in the other, see which fills up first".

Not exactly accurate. I've completed four AM5 builds in the past several months, all with four fully populated RAM slots, and all running at the modules' rated EXPO speeds. All of them passed extensive testing in the shop before being put in service, and all are still in service right now running perfectly as expected. Three are running 64G in four 16G modules. Two of these are primarily flight sim machines (to include DCS). In my experience, a successful outcome depends a lot on the exact components used, and also to some extent who does the work.

Imagine that. Clear, first hand evidence that better drive performance absolutely can result in better gaming experience - and that's even when coming from one of Samsung's own very capable PCIe 3.0 drives. Let the haters and naysayers go on all they want about how one PCIe 3.0 drive is good enough for the entire system (OS and games), and how faster drives do nothing for gaming performance aside from "faster load times". After all, they've been saying that for decades now, even as both empirical and anecdotal evidence show otherwise. There is an ever-growing body of substantial evidence to show that is an inaccurate assessment. While more conservative configurations (i.e. single and/or slower drives) may be considered adequate (by some), they simply cannot yield the increased performance that multiple and/or faster drives can. This is due to physical factors related to the various drive types/arrangement, which are well-documented for decades now.

Absolutely this.

Again, all of this is easily resolved by working with someone who is experienced and trained. And no, it doesn't necessarily have to cost anything, either. (And I might add that, if they're interested in learning about it, that opportunity usually exists and is well worth the effort). I'm pretty sure the numerous DCS players I've already done this with would provide very positive references. However, I don't want to get too far into it, because next thing you know I'm being accused of spamming the forum to try and sell something - which is total b*llsh*t, but there it is.

Model numbers (from System Information) will almost always tell you all that by doing a little legwork online. There's no shortage of information that's available. The issue is knowing what the information is good for.

The post above I as referring to wasn't yours. There was a post before your saying neither CPUZ nor System Info (not System>About) provided drive details, which is not accurate. Sorry for any confusion. But I think part of it is that you're referring to System>About as System Information, which it's not (see the title bar in the pic I posted earlier, that is System Information...which as you noted can be searched for but doesn't require knowing anything about MSINFO32). FWIW I wouldn't refer to System > About as an applet, and I think the actual name is "About your PC", as stupid as that is. We can thank Microsoft for the hot mess that Windows has become. I should add here that no amount of system info, applet or even utility will tell a typical user what the information actually means. Anyone can run these apps/utils etc. but most forum 'experts' don't really understand what the information is good for. There are countless examples of that posted here all the time. My advice is, and will remain, to seek the assistance of a qualified professional.

Windows System Info shows BIOS version. Next to last line in image above. Also, unlike what the prior post above says, System Info does show drive models (as shown below). You can look up the model online easily and get all the specs you want. Unfortunately, as I mentioned above, System Info cannot give you memory timing details - and yes, it matters. I've seen plenty of times the memory was configured improperly and not running as it should, or mixed modules (never a good idea) or a host of other issues. CPU-Z does provide this.

I stand corrected. The 7B78 is a model number, which in haste was confused with a BIOS version. Prior post edited.

If someone has a setup that already has 3600 RAM on it, and can offer a better deal than buying stuff piecemeal, how's it better? @ Rayzer You have a MSI X470 Gaming Pro Carbon. It looks like the BIOS is current with that required to run a 5800X3D CPU. So that's pretty much a 'drop in' upgrade - but it absolutely depends on whether you're comfortable doing that sort of thing. I do strongly recommend seeking professional help if you're not certain.

Everyone's entitled to their own opinion. Fact is another matter. And speculation is pointless.

I think it's widely accepted that the Zen4 CPUs work best with 3600 CAS 16 modules. Besides that, he may not even need new RAM. Maybe he's already got something decent. No point in tossing it out just because System Info doesn't tell you anything about timings etc. Again, just a few points... As for platform upgrades, etc....with the right advice and qualified help, those things are all addressed.

Suppose someone's RAM is not setup/running properly or they've mixed modules, etc? Seen that many times (or any of a number of other problems)...and again, these are also just examples. I don't think System Info will tell you those things. Still...no one's saying System Info is useless, relax. But it is obvious it doesn't report some useful info. (edited, removed comment re: model number)

Really? What I posted above (BIOS version and motherboard model) were only examples...does System Info show memory timings etc like CPU-Z does (both what is currently running *and* the Serial Presence Detect JEDEC/XMP/EXPO values)?

CPU-Z offers far more usable info than Windows System info - not the least of which is you can get the motherboard model and BIOS version, which in turn allows identifying whether/what upgrades are possible.

I think it's accurate to say using a TrackIR-based system would require significantly less of a system/upgrade than a VR-bsed system would, generally and broadly. On the Ryzen CPU: Easy There's a nifty utility called CPU-Z you can download and it will tell you what model etc. It's free, small, legit, and easy to install/use. TBH it's more or less a basic staple on any machine I've built and very helpful to have: https://www.cpuid.com/softwares/cpu-z.html Looks like this when you run it: