Makes sense for AMD to make them as any cpus that have a defective core end up as the 7900x3D.
-
Competitor rules
Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.
Poll: Ryzen 7950X3D, 7900X3D, 7800X3D
- Thread starter kindai
- Start date
Associate
- Joined
- 19 Nov 2015
- Posts
- 108
Do we know for certain that the 7900x3D is split 6/6? Could it be 8/4? I appreciate that makes the yield equation a little odd but there's no stopping that being the case, right? The clocks do appear to suggest 6 core vs. 8 core.
I took a look through the AMD announcement videos and the AMD website but couldn't find it confirmed. Its been reported through multiple outlets that it would be a 6/6 split https://www.digitaltrends.com/computing/what-is-amd-3d-v-cache/
Deleted member 258511
D
Deleted member 258511
What people need to remember is, all the gaming benchmarks will be done vs a stock Intel CPU. If you have your 13900k tuned and fast memory tuned also, sometimes 8000mhz and above, I can’t see the 3d keeping up. Yes at stock for stock maybe in some titles but not when you unlock the full potential of day a 13900k with some good memory.
Caporegime
I whatched JayZ2Cents test a $75 air cooler on his 10900K test rig last night, it was as he said tuned, he had been using a 360mm AIO on it in an open test bed, with the $75 air cooler on it the thing would freeze up with in seconds on every R23 run and long before it reached TJmax temps, he had to back off his overclocks to complete a 20 second R23 run, which he excused as a difficult thing for a CPU to do, no its not.... its about as easy as it gets, he wants to try baking lighting in Unreal Engine for 20 seconds which already is very much harder never mind 9 hours straight, my CPU does that with ease and never lets me down....
My point rather being pushing CPU's to the absolute limit with the most expensive gear you can find for a 20 second R23 run is a ##### huge lie.
Last edited:
Technically, the workloads for calculating light maps and ray tracing will be very similar. The real difference is the length of time it takes, 20s vs 9 hours at full load, just saying
. Also, Unreal Engine will use insane amounts of RAM too.
Last edited:
Caporegime
Technically, the workloads for calculating light maps and ray tracing will be very similar. The real difference is the length of time it takes, 20s vs 9 hours at full load, just saying
Its basically the same thing, the difference is the scene in R23 is 6X6m and has a couple of thousand path traces, my map is 4KM X 4KM and has billions of path traces, the floating point calc and memory trace switching is insane by comparison, it puts the transistors under extreme stress and increases the chances of an error in comparison by a factor of millions.
All i'm saying is these people use the most expensive gear and then use that to push the hardware right to the brink of where it can just complete an R23 run and call it stable, as if that's what you should expect with your not so expensive ancillary components running it like that day in day out for almost any task, its complete BS.
Last edited:
Deleted member 258511
D
Deleted member 258511
That’s why I don’t watch his channel. I would only use Cinebench for a baseline only. It’s like memory tuning, some people think if it boots and runs Aida64 and a quick run of OCCT it’s stable. It isn’t. I’ve heard of people changing only the primary timings and wonder why there is little to no difference in performance.
What I’m saying is a 13900k at stock is showing that CPU at its very worst. That is what will happen in the mainstream anyway and people will go with that.
Caporegime
These people are clueless, for sure.
Associate
Ironic you mention Elden Ring as it requires a 6 core CPU per official minimum system requirements and it being a simple game.
Though wait it actually requires only a 4 core CPU Ryzen 3 3300X with SMT. Though non-SMT 6 cores i5 8400. Wonder why no Intel 7700 or 6700K on there.
Requirements Test
Check the system requirements. Can I Run it? Test your specs and rate your gaming PC.
www.systemrequirementslab.com
Yet all the talk that Cyber Punk is so intensive and some say it can even benefit form more than 8 cores??
Look at its minimum and even recommended system requirements:
Requirements Test
Check the system requirements. Can I Run it? Test your specs and rate your gaming PC.
www.systemrequirementslab.com
Its recommended do not even have more than a 4 core CPU lol.
Are the manufacturers way off, or do games still run well with even only 4 cores?
Even Far Cry 6 which is very CPU intensive only has minimum requirement of 4 cores and recommended 6 core 3600X or 8 core non-HT Core i7 9700
Requirements Test
Check the system requirements. Can I Run it? Test your specs and rate your gaming PC.
www.systemrequirementslab.com
Last edited:
Associate
Despite Intel troubles their P cores still beat AMD's Zen 4 cores in overall IPC and can clock higher. Though AMD at least has the foundry to go beyond 8 good cores and has no peasant cores.
other advancements exist , look at the 5800x3d has lower clocks and uses ddr4 yet in some games it can even be better ?? will be interesting how the zen4 x3d models hold up overall in number of different games inc lows
Last edited:
Caporegime
Those cores are 3X the size and use 2X as much power for 10% better IPC.
They are a very long way from good.
Last edited:
Caporegime
I'll add to it.... when you're talking about architectural efficiency, the Zen 4 cores are very much closer in size and power to the E cores than the P cores.
Last edited:
Associate
Some truth to that though they can clock higher as well and have 20% better IPC than Zen 3 and like 8% better IPC than Zen 4 plus better latency. Plus clocking higher better performance overall to add on top of better IPC
Though you are right about being 3X the size and using more power though it is not near twice power usage. More like 20 to 30% more.
Though certainly much much better for Intel now than the days from September 2003 until July 2006 where single and dual core Netburst based Pentium 4s were space heaters with much more power usage compared to single and dual core Athlon 64 which used less power and spanked Pentium 4 in like almost everything even clocked over 1000MHz lower as its IPC was like 70% better.
Intel in far better position now even if they use a larger sized CPU and takes more power at least they actually have better IPC unlike the disaster the were compared to K8/Athlon 64 then before they came out with Conroe which flip flopped everything and sent AMD reeling and always behind and usually dramatically so in almost 13 years until Zen 2 release which brought competition again kind of like Athlon XP and Pentium 4.
Last edited:
Reviews come out day before release so you have enough time to decide
Deleted member 258511
D
Deleted member 258511
I honestly don’t see why companies don’t if they are as good as they say they are. Same thing with games, they get all their pre orders in and show no real game footage, which let’s face it almost always means they’re *****.
Not saying this is the case here, but they’ve given people nothing but dates being pulled and unlocked but not really etc.
They all do it intel , Nvidia, AMD reviews day before release regardless if the product is good or not
Caporegime
Intel have improved a lot, i've recommended 13'th gen over Zen 4 on this forum, i like them, but they are a brute force approach, that makes them expensive, for Intel that is, margins even on these retail CPU's is very low if anything at all, this is with Intel forcing low margins on retailers, who are recommended Zen 4 purely because the margins they get on those are 4X higher.
That's retail, Datacentre the situation is frankly dire.
Last edited: