Instalează Steam
conectare
|
limbă
简体中文 (chineză simplificată)
繁體中文 (chineză tradițională)
日本語 (japoneză)
한국어 (coreeană)
ไทย (thailandeză)
български (bulgară)
Čeština (cehă)
Dansk (daneză)
Deutsch (germană)
English (engleză)
Español - España (spaniolă - Spania)
Español - Latinoamérica (spaniolă - America Latină)
Ελληνικά (greacă)
Français (franceză)
Italiano (italiană)
Bahasa Indonesia (indoneziană)
Magyar (maghiară)
Nederlands (neerlandeză)
Norsk (norvegiană)
Polski (poloneză)
Português (portugheză - Portugalia)
Português - Brasil (portugheză - Brazilia)
Русский (rusă)
Suomi (finlandeză)
Svenska (suedeză)
Türkçe (turcă)
Tiếng Việt (vietnameză)
Українська (ucraineană)
Raportează o problemă de traducere
Reguli și ghiduri pentru discuții
Raportează această postare
No.
The connection on regular 4-DIMM motherboards is slightly slower than on 2-DIMM-only motherboards, even if in both cases only 2 RAM sticks are inserted. This is due to the daisy chain connection with 4 slots.
Intel's official specs are allegedly for 2-DIMM-only motherboards and not for regular motherboards.
At least according to Wendell from the video.
Google Asus ROG Maximus Z790 Apex used by Intel to see the difference. Such motherboards can achieve higher RAM speeds than 4dimms boards.
If you're buying a flagship i9 but limiting the system to ~5600 MHz DDR5, you're doing it wrong and undercutting performance by a potentially large margin depending on the load. So memory specification means very little if anything at all.
AMD on the other hand show their benchmarks with EXPO profile enabled (6000 CL30).
Just to be fair to both sides.
and there are ram kits slower than 6000 cl30
amd cannot guarantee that it will work in all situations or with 4 dimms
To me, this whole thing just looks like extra ammunition to fire at Intel when it doesn't really matter and could again just be precipitation from already known issues. And not everything that these guys claim is the complete truth, it's mostly a matter of professional opinion.
https://www.intel.com/content/www/us/en/products/sku/236773/intel-core-i9-processor-14900k-36m-cache-up-to-6-00-ghz/specifications.html
So it's not clear under what conditions these are the supported speeds, but a pair of single rank DIMMs is about as light as it gets, so it's fair to assume this would refer to that.
So where is Intel stating they only support up to 4,400 MHz as a maximum (and not just under some specific conditions) for two DIMMs? Because unless there's a contradiction from Intel here, then there's nothing to this.
Maybe that speed drops with dual rank DIMMs, but Intel does state "up to" which doesn't mean 'always". And usually, these speeds Intel and AMD give are conservative anyway.
AMD's page, at least for the 7000 series (unless all have been updated recently?) is actually more clear with regards to this, as it states this...
"Max Memory Speed
2x1R DDR5-5200
2x2R DDR5-5200
4x1R DDR5-3600
4x2R DDR5-3600"
https://www.amd.com/en/products/processors/desktops/ryzen/7000-series/amd-ryzen-9-7950x.html
Interestingly, it isn't listed as dropping with dual rank DIMMs for AMD (there's a huge hit using four DIMMs instead of two DIMMs though), but in reality, dual rank DIMMs will likely still lower it. AMD is simply listing a more conservative number for the single rank DIMMs.
With just a pair of single rank (and maybe even dual rank) DIMMs, you'll probably get 6,000 MHz (if not higher) on both AMD and Intel.
Can you clarify this part? Specifically, when using two DIMMs on a four slot motherboard, where are the DIMMs populated?
Because the only thing I can think of that would reduce the frequency you can stabilize with two DIMMs on a four slot motherboard versus a two slot motherboard would be if you put the two DIMMs into slots 1 and 3, as opposed to slots 2 and 4 like every motherboard manufacturer recommends, and for this very reason.
This is not exclusive to Intel and applies to AMD as well. With daisy chain topology motherboards, which motherboard manufacturers have predominantly moved to, putting only two DIMMs into slots 1 and 3 will leave slots 2 and 4 as "open ends" and introduces electrical interference, which can limit the maximum frequency you can stabilize. You avoid this by using slots 2 and 4 (or by using a board with only two DIMMs, or by populating all DIMM slots, but then that last one also reduces the maximum frequency you can stabilize anyway).
I strongly advise you to go over the link I provided in my first reply (the one on the MSI forums) as it really covers everything this thread pertains to. In short, it's "heavier memory configurations are harder to stabilize, and thus have lower frequency ceilings, and thus have lower officially supported speeds in some conditions".
I also refer to BuildZoid who also talk how two DIMMS slots only motherboards are more stable with higher speed memory.
Here is the same video again for convenience.
https://youtu.be/Q8XLvZTqyX8?si=He1vv4yw4_B_b1M5
Actually I have been wondering if ram speeds of 6000 and higher are a contributing factor in the soup of things that affect the current issue.
Nevermind that last bit. Apparently there's optane D.D.R. 4 sticks.[www.intel.com]
You're the one making a thread, and you're making claims that Intel is being misleading and that it is a bombshell. So being unable to clarify it seems odd to me.
If you made the thread saying "what does this mean", it would be one thing, but you seem to implying things with it, so being unable to clarify it means you might not understand it to me.
So I watched for a few minutes after 14:18 in the original video and here's what Wendell says... (I may be paraphrasing a bit between using closed captions and trying to get what he says exact, but I'm repeating what I hear him say without trying to change the meaning).
"And so you get into looking at the memory thing and it's like wait a minute, in the fine print in the memory spec it says that if the memory is routed, if you have a motherboard that has two DIMMs per slot routed, which is basically all of them except one or two, the only DDR5 memory you can expect is like 4,400 MHz, and so if your CPU happens not to work at a memory speed above 4,400 MHz, that's your tough luck, and so wait a minute, that's a little bit of a bombshell if you think about it because Anandtech, they did a lot of their testing at 5,600 MHz which is the highest supported speed, but you only get that speed if you have a situation unattainable in the real world, which is you have a motherboard with only two DIMM slots."
So there's a bit to unpack here.
The part where he talks about if the motherboard is routed with more than two DIMMs per slot would be referring to any motherboard with more than one DIMM slot per channel. Since consumer boards are dual channel, then yes this would mean any four DIMM slot boards.
He's claiming that the maximum supported speed is 4,400 MHz instead of 5,600 MHz then. Okay?
My question is this. Where is the contradiction from Intel? The 5,600 MHz speed is listed as "up to" on their website, but with everything shady Intel is doing lately, I do consider maybe this was stealth changed, or maybe they have contradictory information in different places. I'd be open to seeing it, but at the same time, it... doesn't matter.
His bigger issue with this seems to be that outlets (he lists Anandtech in particular) reviewed the performance at 5,600 MHz, but Intel only supports 4,400 MHz.
Well... you might want to look at the entire review landscape because that's happening all over. I guarantee you a lot of testing on AM5 is done with 6,000 MHz, for example, but look at the numbers I listed above that AMD themselves state they support; they're not the 6,000 MHz that AMD's CPUs are often tested at either.
So this wouldn't even be on Intel, but on the review outlets. Well... the reason they review them higher is because these higher speeds are usually going to work. Especially for gamers dropping only two DIMMs of lower capacities in (in this context, "lower capacities" would be single rank stuff, so that would be a pair of DIMMs that are 16 GB or 24 GB each on DDR5 right now).
You really start running into this problem with higher rank DIMMs, or four DIMMs over two, or using two DIMMs in a four slot motherboard in slots 1 and 3 instead of 2 and 4, the latter of which has an easy solution.
Nothing he says is wrong... well other than the part about it being a bombshell, because unfortunately this is anything but rare. You seem to be taking another meaning from it, perhaps because of his choice of using the word "bombshell" there.
We've been "reminded" of this time and again. Profile speeds are often "overclocking" and not guaranteed. If you're sticking to "gamer capacities" and two DIMMs, you'll 99.9% of the time be fine with whatever "sweet spot" speeds reviewers test at. It's higher capacity configurations, such as ones that need four DIMMs or dual rank DIMMs (or both) to hit their capoacity target, that you need to start being aware of the fact that you might no longer be able to hit your desired "sweet spot" frequency.
Again, please, the thread on the MSI forum that I linked to described everything relevant about all of this.
I am running four dual rank DIMMs myself. This is on DDR4 and I'm running at 3,600 MHz, so I'm hitting the "sweet spot" for the latter AM4 CPUs. It's a very heavy configuration, but I'm hitting it. Albeit, I'm using later DDR4 motherboards, BIOS, and RAM. This stuff matures as it ages. Remember when AM4 was newer? People were struggling to hit 3,200 MHz with four DIMMs or maybe even two dual rank DIMMs back then too.
If I wanted to move to AM5/DDR5 right now, I likely would not hit the sweet spot frequency if I went with four DIMMs. The best compromise I'd be able to make to go up in capacity and still have a reasonable chance of hitting 6,000 MHz would be to use 2x 48 GB DIMMs, which are dual rank, for a total of 96 GB. 128 GB+ would need four, dual rank DIMMs until 64 GB consumer DIMMs release in wide availability. And using four dual rank DIMMs, from what I can find using anecdote from most on AM5, might mean needing to drop to 4,800 MHz or 5,200 MHz. I'm not saying 6,00 MHz would be impossible, maybe someone has done it, but it certainly wouldn't be likely and near guaranteed or plug and play right now.
Bullzoid is an extreme overclocker from what I know, so he falls into one of thew two camps I mentioned (high capacity users, high overclockers/tuners). He certainly knows his stuff, but saying "a memory configuration is lighter with less DIMMs" is a bit of a water is wet thing, so you'll have to specify what, exactly, he is saying and what it means for me to give more of a reply. As it stands, there doesn't seem to be anything to this. Perhaps at the very highest levels (where Bullzoid is), and with DDR5 especially, the extra two DIMM slots being cut entirely, or moved closer to the CPU, help stabilize higher speeds. Again, that's a "water is wet" thing, and is also unlikely to realistically impact most users. Why do you think the industry is looking to move to CAMM2? It helps stabilize 8,x00+ MHz speeds.
We're simply at the point where it's getting harder to stabilize higher frequencies. What both of them are (likely) talking about falls entirely under that.
Wendel looks surprised and disappointed about the supported RAM speed on Intel. According to him, AMD may actually have better specs for 2 RAM sticks on 4 DIMMs motherboards. He talks about this at around 24:40. He says that AMD is honest about their RAM speeds when Intel small print is not. Maybe this information is as obvious as water being wet for you, but for me, learning that Intel supported RAM speed is 4000-4400 MT/s is wild.
Sure, almost everyone uses EXPO/XMP anyway, but that's not the point. 14900K based gaming servers weren't stable with 5600 MT/s RAM, but they were stable at 4400 MT/s. This seems to be unrelated to the degradation issue.