Jump to content

kksnowbear

Members
  • Posts

    881
  • Joined

  • Last visited

Everything posted by kksnowbear

  1. Holy sh*t I actually LOLd and I've not heard a more eloquent summary of this mess since it started. 10 year old turds....indeed LOL
  2. I would submit there's a lot more going on there than has anything directly to do with CPU performance. Just as an example, a guy can be a complete moron and still have a YouTube Channel, and if enough people watch it, he can get sponsored and/or get most if not all his hardware for free. I know someone who streams and did exactly that, and I'm sure it's not uncommon at all. This is definitely among the reasons I mention that having all the nice hardware doesn't make you an expert. Trust me, I've seen plenty right here on this and other flight sim forums to know. Sad part to me is that a lot of people (by no fault of their own) can't tell the difference... ...long and short: Again, be selective about who you work with and take advice from. Anyone worth a flip will have no difficulty proving they've done it for others who are satisfied.
  3. How very interesting. I have not less than three 12xxx builds in my shop at this very moment. And for the very reasons cited: Less cost, and no funny business. FWIW I also agree generally about both low- and super high-end motherboard models. Yet another reason I will continue to say work with an experienced builder who can show demonstrated success. There are instances/people where some of the high-end boards' features are appropriate and useful (my very own X670E was chosen specifically for two such features)...but the best way to know is deal with someone experienced and knowledgeable.
  4. No worries, glad if it helps. Please feel free to contact me if you wish.
  5. The Z boards are the top-end, performance oriented, "enthusiast" level boards, yes. (Z comes from the chipset, and generally you don't find Z790 chipsets on entry level boards. That said, there are better and worse Z790 boards as well.) That said, I'd avoid Intel like the plague right now, seriously. The BIOS updates might mitigate the problems to some degree, but it won't solve them. In order to effectively reduce the risk, you'll still need steps like @LucShep has outlined. And TBH it's really still mitigation, not resolution. IMHO and FWIW as always
  6. This doesn't exactly strike me as a build that's being put together with budget as a priority. What I mean is that there's a lot of cost here for fairly high-end stuff, so I am assuming this isn't about doing things economically. (Yes, I agree) a 7950X is probably money wasted on the "top end" when there's nothing at all "lacking" about a 7800X3D. Also, while I understand sales and promos can often mean it makes more sense to buy high-end components, I really hate wasting money on cases. It's basically a steel box. If it's well ventilated, roomy enough, and (if desired) has a transparent side panel, then it's adequate. I flatly refuse to pay $100+ to get those few features, and it's absolutely not necessary to do so. That's more about someone's 'part picker' list trying to upsell than anything else. Sounds cards are completely unnecessary these days in all but the most exceptionally rare situations (audio engineers, but they don't really use PCs anyway...). Not a fan of Corsair RGB RAM (or anything else) because they want to force you into using their RGB bloatware, which (of course) doesn't play nicely with others' (including motherboards and GPUs). If you really desire RGB (a matter of preference) then there are much better ways to go. I also agree that air cooling is adequate and can be less expensive. That said, I use liquid cooling myself and there's nothing wrong with it, necessarily - it works better for cooling of course, though it is true to say it's not as reliable over time as air. I'm not a huge fan (no pun intended) of 'tower' air coolers for one simple reason: The airflow is parallel to the motherboard, meaning there's no cooling effect on critical motherboard components by residual airflow off the CPU cooler. Perhaps the only thing both Intel and AMD got right in stock cooler designs is that the fan being "flat" on the heatsink, blowing air perpendicular to the board, also directs airflow at VRMs and RAM. Tower coolers, by design, cannot do this. So if you use a tower cooler, then it's all the more crucial to make sure you have adequate airflow directed at those components *and* not obstructed by other parts.
  7. This happens a lot, unfortunately. Problem is, builders/sellers who are untrustworthy and/or lack knowledge ("builders" in this context to include DIYers.). They either don't know what they're doing outright, or they take advantage of buyers who aren't knowledgeable. Numbers like "14900k" are very common and recognized even by buyers who don't have a clue. So, unscrupulous sellers match top-end CPUs (because the buyer recognizes the number) with a cheap crappy board (because buyers don't know/don't care/don't try to educate themselves). BTW: pre-built "PC Vendors" like iBuyPower and Alienware do crap like this often, and yet still charge what might be appropriate to a complete build of all top-end components. Cheap, feature-stripped (and sometimes BIOS-hobbled) motherboards, power supplies, memory modules, even GPUs (a cheap a$$ low-end 4090 can still be advertised as a "4090"). I've seen tons of >95w CPUs on boards that won't support anything over ~90w. You can see it all over this forum in sigs, people who (for "bragging rights") bought the most expensive top end CPU on the market, then paired it with a stripped down board that will never fully support the CPU's feature set, because they have money and think they know what they're doing. I'll keep saying it: There is a huge difference in a professional computer expert, and forum geniuses with lots of money to buy all the latest top end hardware (or else buy a pre-built), then represent themselves as "experts" simply because "I have a 4090 and a 4k monitor, so I must know what I'm talking about". Building a few computers over your lifetime doesn't make you an expert, nor does throwing money at expensive hardware you truly don't understand (whether you assemble it yourself or pay someone else to). Find an experienced, trustworthy builder with verifiable references, who guarantees customer satisfaction. Period. (It doesn't necessarily have to cost you anything, and working with the right builder can often help save you money, especially over time).
  8. The BIOS save issue you discussed yes, is stupid...been that way forever, far as I recall. It's obviously not intended for use with different BIOS versions; I get the impression it's only a sort of backup mechanism in case your BIOS becomes corrupted (as opposed to a recognized version change?). TBH although it does seem stupid, I can see there are reasons for it being that way - for example, can't have some setting backed up/restored that they removed from the newer BIOS version. (This is a variant of what we in the field used to call "Versionitis") On the MCR...there are TWO settings. Did you do both? I think your crashes/BSODs are possibly because you only changed MCR. Both Memory Context Restore and Power Down Enable must be enabled (in my Asus board's BIOS that's what they're called anyhow) HTH (PS A 34-second boot time these days is pretty much normal, at least from what I see...sometimes it goes much faster, but if we're including 'to the desktop' then ~30 seconds isn't bad at all. When the memory 'training' decides to show out, it can take several minutes - it is very obviously different).
  9. Yeah, I usually bump up the RAM voltage just a tad, mostly because I'm almost always using full banks. This has always helped to avoid problems in these configurations (this includes four recent AM5 builds).
  10. Perhaps the OP would be good enough to favor us with a pic of the 2080Ti in the case that the original pic shows? (And yes, I would agree the PSU is mounted upside down...I noticed this before and forgot to inquire. I'm not sure anybody made a case with a bottom-mount PSU that has no intake grill/vent underneath - could be wrong, but I don't think I've ever seen one).
  11. Well...of course, it has to do with "features" - that is, what features do you want/need vs budget/cost. For example, a board might have every conceivable feature you'd ever need, but cost twice as much as you intend to spend, so it's not a good choice. For the 7000 CPUs ('Zen4') there are four types: B650, B650E, X670 and X670E (in order from least costly/fewest features to most costly/featured; there is also an entry level A620; not at all meant for gaming/performance, thus we're not discussing that lol). Of the four we're talking about, the basic difference is PCIe 5.0 support/number of lanes. We're still a long way from saturating 16 lanes of PCIe 4.0 bandwidth, so some people say there's no need to pay for PCIe 5.0. However, things do change, they are changing, and the rate of change sometimes accelerates once a given feature 'catches on' (note it can also fizzle out and disappear, for better or worse, like eSATA, Firewire, SLI, etc). I think that it's not GPUs that will really benefit from PCIe 5.0, I believe it's storage. Without getting too far into it, even though most other parts of a PC have come a long way in terms of speed/performance, there are still people using machines with SATA SSDs or even older mechanical hard drives. More recently, the hardware has finally advanced through different levels of PCIe such that PCIe 4 devices are running up to approximately 12 times the speed of a SATA SSD, and PCIe 5 can reach past 26 times SATA SSD speeds. The problem - as is often the case, is that now the hardware has exceeded what the software was really intended to do. So there's not really much taking advantage of the tremendous advances in storage speeds to improve game performance, beyond simple faster loading times. Faster loading times are OK, but that increased storage performance can now actually improve in-game performance. Yes: Frame rates can actually improve with faster storage, and this can help improve texture loading, not just maps/levels, so that moving through game 'worlds' is smoother. Again, I don't want to get too far off course on it, but it is real, it is already happening, and IMO it's going to 'catch on' to the point support will be standard in all games. Due to things like the pandemic and other factors, it's been slow to gain ground, but I think it'll grow exponentially. More significantly, it represents the first time (that I can think of) where storage performance can actually improve game performance, rather than hurt it. So, PCIe 5 - IMO - is an important feature in newer CPUs, chipsets, and boards. GPUs may not really 'need' this bandwidth now, but storage can definitely benefit. Here's a chart showing how each of the chipsets/boards implements PCIe 5: >AM5 chipsets< (note this chart also shows thr newest 800-series chipsets, but I'm not addressing those here) Here's a really nice Google Docs sheet showing all the various boards from all the major manufacturers, with features broken out: >AM5 board sheet (works best in Chrome)< The four chipsets I mentioned before (B650, B650E, X670 and X670E) will - broadly and generally - support a greater number of PCIe lanes, and PCIe 5 vs 4, as you go through in the order I've listed them. Generally: X670E will have 5x16 (GPU) and (up to*) two 5.0x4 and two 4.0x4 M2 slots X670 have 4x16 GPU and one 5.0x4 and (up to*) three 4.0x4 M2 slots B650E have 5x16 (GPU) and (USUALLY*) one 5.0x4 and (up to*) three 4.0x4 M2 slots (most have only two 4.0x4) B650 have 4x16 (GPU) and (USUALLY*) one 5.0x4 and (up to*) two 4.0x4 M2 slots (many have only one 4.0x4) *("up to" generally depends on board model and physical size; eg microATX, ITX boards are smaller thus fewer m2 slots) *(USUALLY: MOST B650E boards have one 5.0x4 slot, but a couple top-end models have two. FEW units have three 5.0x4 and one 4.0x4) Please note this is **generally** so don't hammer me if there are exceptions here and there. So you can see that the difference in different boards mostly comes down to PCIe version and number of lanes/available slots. Beyond differences in chipsets, There are other, less obvious features you might want; one example of this is I tend to look for boards that support some type of thermistor (Asus calls it "T-Probe" IIRC, others have their own names}. This specific feature is often only found on higher-end boards, but it's something I've leaned to appreciate and value. I can use the sensor input on the board to physically monitor the GPU temperature, and control fans on my case based on the actual GPU temp. I don't have to have fan monitoring/curve controlling software (like SpeedFan) running. The board I chose also supports a specific type of overclocking which even works on X3D CPUs, something that isn't typically supported without that feature (of course, it costs more). It is *very* rare, even among enthusiast boards like X670E, and even more so on non-enthusiast boards like B650E. I don't think there are any X670 or B650 (non-E) boards that have it. There's also the number and type of USB ports - this matters more to most flight sim players I think than it does to other types of gamers, because of the number of devices some flightsim enthusiasts use. It's a good idea to actually count the number of ports you're using now so you can make sure you're looking at boards that have what you need. So, like usual, budget will probably come first, 'brand loyalty' usually comes next, then figure out which features you want for the cost, and check the sheet to see what your options are. Even B650 boards can be suitable, if you're not going to use a lot of storage and don't care about GPU having PCIe 5. If you want PCIe 5.0 GPU slot, a B650E might be a good choice. If you need PCIe 5.0 graphics *and* want multiple PCIe 5.0 storage slots, you're probably looking at a ATX sized X670E board (there are a couple B650E exceptions with multiple PCIe 5 storage slots). Hopefully that answers the question. Sorry for the length but there is quite a bit to consider. I welcome any questions. Most likely, you're seeing the long boot times due to what's called memory 'training'. I have an Asus X670E board, same thing. I've done several Am5 builds now and they do seem a little 'fussy' about memory, particularly in certain cases. Good news is it's fairly straightforward to address: If you haven't done so, there are two BIOS settings that you can change that will resolve this - although I would say that on *extremely* rare occasion, my board has still decided to take longer to boot than usual. I have done some looking into it, and it does seem a very complex process, so not a huge shock to me TBH. But, as I said, the two BIOS changes all but eliminate it on the units I've built.
  12. There is very little these days by way of practical competition for the 7800X3D where gaming is concerned. The 13th and 14th gen Intels *might have been* an option, spec-wise at least, before the current mess. But I personally couldn't feel confident in them after all that, nor would I consider recommending them in my professional capacity. The next gen Intels might be better...but then again, there's already a set of next gen AM5 CPUs. Just really hard to consider Intel, of late. Also if you intend to use the machine for productivity, a different AM5 might be in order. That would depend on exact workload and of course budget.
  13. Indeed, if you're among those who can reuse storage for now, nothing wrong with waiting out pricing. You're more or less in Case #2 above. I would say that whether you should consider a T700 or 705 in the 'near future' will depend on what it's worth to you (and what 'near future' means to you lol). TBH, if you can find prices like I showed above, once you have the new platform then I can't imagine it's a good idea to go backwards and buy Gen4 storage new when much better performance costs not much more. But of course, that depends on price. I still have two Gen4 drives on my X670E board that I use plenty, in addition to two Gen5 drives. Depends on the usage. I do still maintain that there is already improvement in DCS (and any game) from faster storage, but the 'forum experts' get all in an uproar at the mere suggestion - which I find interesting, as apparently none of them actually have Gen5 drives and thus can't possibly have tested it first hand...and even if they did, their answer will always be it's "not worth it". Factually, what is 'worth it' is entirely and irrefutably up to the guy paying for it. The three Gen5 drives I've owned were absolutely "worth it" to me. If I see a $20 bill laying on the ground and I stop to pick it up, it's "worth it". If Bill Gates stops what he's doing to pick up $20, he's probably wasting money and/or time lol By it's very nature, "worth" is always relative. Those 'experts' typically only consider performance has improved if FPS is higher, which is completely misguided and inaccurate of itself. There is much more to performance in a gaming computer than sheer stupid frame rates.
  14. It's a good question...a little tricky to answer; not quick because there's a lot to consider, but I'll try to make sense. I figured out basically three use cases: 1. If your current platform doesn't support Gen5 *and* you don't plan on changing any time soon, then it doesn't matter. 2. If your current platform doesn't support Gen5 and you *do* plan on changing it, *but* intend to reuse drives you already have, then it doesn't matter. That said, I'd suggest that when you do change, you go with a platform that supports Gen5. You can reuse your current drives, and still upgrade to Gen5 later if you're not ready right away. 3. If your current platform doesn't support Gen5 and you *do* plan on changing it, *and* plan to buy new drive(s) at or about the same time, then - especially given the prices I cited earlier - it would certainly seem to make zero sense to go backwards to Gen4 or anything else. $20 more for a 2TB drive that's 66% faster? No-brainer. (NOTE: The prices at Amazon have since changed [of course]...but they were exactly as I said for a good while around when I posted that info - and, of course, they'll change again. When you're trying to get good deals, you do have to be patient and watchful). In any event, the argument that Gen5 storage costs twice as much as Gen4 (or whatever other ridiculous figure) is also not necessarily true. I can assure you I didn't pay anywhere near twice what (good) Gen4 drive costs for the three Gen5 devices I've owned. Anyway, that should cover the various use-cases. Two where it doesn't matter at all, and one where it wouldn't be smart to buy anything *other* than Gen5 storage if your board supports it. So you do the math. Also: The AM5 CPUs support it anyway, and if you're buying one of them, it would seem a waste not to take advantage of it. As for the actual impact of going to Gen5 from 4 (or 4 from 3): There are storage technologies already being used, in an increasing number of games, that can actually take advantage of Gen5 drives' speeds to improve gaming performance - and I'm NOT just talking about load times, either. It has already been shown that games which support the newer technologies perform better (yes, higher FPS) with faster drives. Gen5 better than Gen4, and Gen4 better than Gen3. Up until not long ago, storage technology just wasn't designed to take advantage of faster drives. This didn't matter for the longest time, because the drives weren't fast enough to matter themselves (think about older, SATA drives - even SSDs). But, more recently, storage technology has jumped from 7x faster, to 12x faster, all the way to 25x faster than SATA. In the past 10 years, storage speeds have nearly doubled, twice (and were already 7x faster than SATA). But the problem was that while the storage was getting faster and faster, it was still being utilized by the system in very inefficient ways - and this is still largely true today. The way storage is used has failed (horribly) to 'keep up' with the speeds that storage is now capable of. As an example: Imagine building a very fast car, many times faster than the car you currently drive. The problem is, if the roads aren't suitable to driving that fast, your new, much faster car is going to be held back. Again, the technology that benefits from faster speeds is already present in a number of current games. So the 'roads' are ready...well, some of them. There's the problem: So far, only a few games support this newer tech; or, in my car example, only a few of the roads have been updated to allow high speed driving. DCS, regretfully, isn't updated yet. But, for a number of reasons, I think it will be. In fact, I think it's reasonable to assert that *all* new games will eventually support this newer tech - but, like with most advances in technology, whether older games support it really depends on a lot of things. But as far as DCS goes I think it'll be supported based on a few things. DLSS and MT: Both these newer methods/technologies have now been incorporated into DCS by ED. They both required a considerable amount of work to put in DCS, and both are still evolving (and will continue to do so). ED has clearly demonstrated a commitment to putting updated features into DCS, and when it comes to performance increases, the players are always on board. Also, If I'm not mistaken, ED has committed to incorporating support for Vulkan - and Nvidia has now built RTXI/O support into Vulkan. I believe it's been (incorrectly) stated elsewhere on this forum that DirectStorage requires DirectX12 and therefore DCS won't have it - but that doesn't seem accurate since 2022 as per https://www.techpowerup.com/301147/nvidia-brings-the-benefits-of-directstorage-1-1-to-vulkan-under-its-rtx-io-brand There is real, concrete benefit to these new storage technologies, and not just faster loading. I think evidence suggests that we'll be looking at some type of implementation in DCS, in the near future. Sorry for the length but I hope that answers your question clearly, and welcome any questions.
  15. Hi, sorry if it wasn't clear: Yes, the 990Pro is Gen4, and it is definitely not faster than the T705 If I've said something to the contrary I'm not aware of it - please point me in that direction. The table is showing how much of each PCIe version's bandwidth is actually used by the drives, as a percentage. It has been stated that one problem with new Gen5 drives is that they aren't using the full bandwidth of the slot, so they're not all that much faster than Gen4 drives. This is completely wrong, and my table was intended to show this with factual data (some people here on this forum seem to only know what they read in reviews, without understanding or testing the actual hardware first hand, as I typically do). Of course, you do have to consider that not all Gen5 drives are the same, any more so than all Gen4 or Gen3 drives. For every PCIe revision, there are drives that perform closer to the max bandwidth of the slot, and those that perform at a lesser speed. Nothing at all new about that, and it absolutely doesn't mean that all Gen5 drives don't 'live up' to the speed of the slot. This was more accurate when Gen5 drives first came out, a fact which so-called "reviewers" just couldn't wait to complain about - but those days are behind us. The table shows how much of the available bandwidth each drive uses - and should illustrate then that the T705 uses just as much of it's (Gen5) bandwidth at 92.06% as most any other drive of their respective, available bandwidths (more, actually) - one exception being a 990 Pro, which uses 94.58% of it's available bandwidth (on a PCIe 4.0 slot). The 990 Pro uses slightly more of it's max available (PCIe 4) bandwidth - but that's still only about half the T705's available PCIe 5 bandwidth. The T705 is still almost twice as fast as the 990 Pro in terms of read speeds. The 990Pro is limited by the Gen4 bus bandwidth to a max of around half the T705. I hope that clarifies things.
  16. Indeed, that would be great. It happens I have hardware such that I can reproduce the OP's situation, so I decided to do a test. What I can say in my limited testing - which was not at all scientific just because of time, etc and thus decidedly 'down and dirty' - is that there was around 20% loss of the gain going to a 2080Ti from a 1070, due to being paired with a platform very similar to the OP's 2600k/Z68/PCIe 2.0.
  17. I mean, I do get it, I love working with older stuff too... But you're not making a 2080Ti do more than it was intended for...due to the old platform, you're actually making it do less. When it comes to the 'hobby' part of computer work. I usually draw the line at where I can do more with older stuff up until the point that newer stuff at comparable/less cost can do the same/better. Either way I am glad you're happy with the outcome
  18. That's the spirit. Just gotta keep digging, be patient, do your legwork, and for Pete's sake ask for/check references. (And FWIW I agree completely, eBay is not usually very impressive...I like to think that's where I come in; used hardware (mostly) without all the drawbacks). I have one of the boards in your link, it is really just about as bottom of the barrel as you can get lol but it would definitely fit the bill for that kind of money. The RAM honestly looks like knock off HyperX Fury modules, seen lots of those but yeah...what speed? 5600G is a fairly good GPU, I have one here and it performs OK. Especially nice if you want to run a second monitor without putting more on your GPU. And it comes with a 1TB drive too! Actually a decent find for tight budget entry level.
  19. Prime95 doesn't have to use AVX instructions, and you don't have to change any files so it won't. It's in the dialogue box for testing, and can be changed every time, without ever editing any file. (What you're talking about is changing the default when it first opens; again, this is not required to change the behavior for testing). The AVX BIOS offset would (likely) be set for zero, if I'm not mistaken, so if that were causing the different clocks, he would know that he changed it. And if he knows that, well...he probably knows how the rest works (granted I'm assuming). The BIOS manual doesn't appear to specify, so without the board in front of me, I couldn't say....hence: I say most likely and I believe not because I think I'm wrong but because I haven't specifically done things like looking at his motherboard BIOS (as one example). I don't necessarily have time to rtfm for everyone who hasn't Don't get so excited.
  20. P95 is configurable to not use AVX (in every version I've used, anyway). Unless I'm mistaken, the default in BIOS is usually zero for the AVX offset, so unless you change it, it doesn't lower AVX clocks...I could be wrong here though.
  21. While I don't disagree concerning used parts, the key is diligence - as applies to most anything in life. Also, I think your assessment "half as good' is not completely fair or accurate. For a good deal less in cost, I could put together something that's actually much closer than half the performance of what you're suggesting. I won't argue whether used is 'as good' a deal as new (or vice versa), it can go either way. Simply depends on the deal you can find. Again, diligence. In my work, I make a point of showing clients actual prices for new stuff they can go get themselves, with comparisons to what I can offer. I also frequently do the legwork myself to show current used prices (as with eBay, etc). Further, I will usually offer to just provide labor - if they want to buy their own parts, I'll do the assembly at a flat rate, so they can choose whatever parts they want, at whatever cost. (To be accurate, I don't do this with someone else's used parts, only stuff purchased new by the client first hand - though I will also usually offer to do the shopping as well). Something else your breakdown doesn't consider is that when I sell systems, labor for future upgrades (or warranty repairs) is totally free, and I also take used parts in trade - so the cost of any future work is going to be less than just buying new parts again or even used parts from a random source. I value repeat business, and I prove it. The decision is always theirs, and while they don't always choose my offer, the vast majority does. There are times, although seldom, where I just cannot beat another deal - in which case I will advise the client "Go buy that" (and yes, I've done it, more than once). Interestingly, on several occasions, I've had people who *didn't* choose my offer come back when it was time to do a new build/upgrade - and most often, they indicate remorse over not having taken my offer to begin with.
  22. I have to agree, certainly in principle...but I would like to offer some other thoughts. For example: I'm not sure I understand. If replacing the board would be an easy swap, and since you'd have to change the CPU itself (second-gen CPUs don't support PCIe 3.0 even on boards that do)...then it doesn't seem as if it would be a lot more difficult to swap the board/CPU with a newer setup that not only properly supports PCIe 3.0 but also is closer to being matched with GPUs like a 2080Ti. Can you please help me understand why it would be easy to go to a Gen3 platform, where previously you indicated the platform 'can't be changed'? (Do you already have a Gen3 setup available at no cost?) Also, I believe @Aapje makes a point, but I also believe there could be a compromise that would actually be somewhat less expensive. The costs he lists (for a good setup, no doubt) are all for new hardware. Unless you do have a Gen3 setup available at no cost, then I assume there would be at least some outlay for that...and if, as above, you consider (a Gen3 swap) 'easy', and if your situation is budget constrained...then I'd say you should consider saving whatever you can, while trying to find a reputable dealer in used parts. I'm reasonably confident you could find something for well below the $400 that new stuff would cost, yet still be way ahead of where you are now. I should also try to clarify something I said previously. It is absolutely true that - *if* a GPU is otherwise fast enough - running it at Gen2 instead of Gen3 PCIe will cause some loss of performance. The faster the GPU, the more the loss. With that said, however, and as I mentioned, it's not an issue with lesser cards (like your 1070) because they're not fast enough to be 'choked' by the slower bus. The actual effect starts in the neighborhood of a 1080Ti or so, and will of course be more pronounced for faster cards (like a 2080Ti). The thing is, this limitation is in proportion to the speed of the card, so that the effect is smaller at the onset (if you will)...thus cards at the lower end of the 'affected range' won't lose as much as higher-end cards will. What I said is still true: You'll lose some for a 1080Ti, more so for a 2080Ti. But we're talking a smallish loss...it's just that, in this case, it's furthering the loss from running on an older platform. The Gen 2 vs 3 part isn't making as much difference by far as the older platform components are; it's still a loss of itself but it's the smaller part of the overall loss. Going to a Gen3 platform will help - but (obviously) that depends on which Gen3 platform and what you put in it. For example: The third-gen 'Core i' CPUs (3770k, as one example) can still work on most Z68 boards that run second-gen CPUs like your 2600k. And they support PCIe 3.0 (most of them anyway, and provided the board supports it). But the brutal reality is that a 3770k is just not a huge upgrade from a 2600k (maybe 5%??). So it's entirely possible that even a Gen 3 platform will *still* be poorly matched to a 1080Ti/2080Ti. If, by comparison, you went to a much later Gen 3 platform....let's say something like a Z390 (this is strictly for the sake of discussion, mind you). Put a 9900K on that Z390 board, fit it with some fast, low CAS RAM... now we're getting to a platform that is somewhat reasonably matched with a 2080Ti. So, to summarize: Going to Gen3 isn't necessarily going to make a huge change here; it would depend entirely on which Gen3 - and that's a gigantic range of hardware spanning almost a decade of CPU evolution (~Ivy Bridge intro c2012 til Comet Lake was discontinued in 2020 - for the Intel side, of course). Sorry for the length, but I do sincerely hope all that makes sense.
  23. I believe it's possible that the option to set All Cores isn't "exposed" unless/until you select one or another form of overclocking (i.e. XMP or Manual, but sometimes "Expert" or similar). Then you see All Core or Per Core. That's generally speaking, and I would also caution I am not intimately familiar with the Gigabyte BIOS at all. I'm not where I can really look into it ATM.
  24. Most likely the two utilities are just using different types of loads (for example AVX vs not). I believe you can configure, at least to some extent, the testing that P95 does. Don't know about CPUZ stress test as I don't use CPUZ for that.
×
×
  • Create New...