• 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

Will you be purchasing the 7800X3D on the 6th?

  • Total voters
    191
  • Poll closed .
KompuKare said:
If Intel had the packaging technology, a massive cached P Cores only processor would be interesting, though.
Click to expand...
Isn't that the 144 core Sierra Forest?
Gerard said:
I'm also sure they stated they tested vcache on multi ccd cpus internally and the results weren't much better than what they released with the 7950X3D/7900X3D vs the additional cost of adding vcache to the other ccds.
Click to expand...
Or business. Why sell one 2 CCDx3D chip when you can sell two 1 CCDx3D chips? Plus the infinity fabric latency is still there. If Zen 5 has 16-core CCDs then there's probably no need to ever sell 2 CCDx3D chip. Plus the 7800x3D is a worse bin than the 7950x3D. I wouldn't be surprised if some are failed 7950s and the non v-cache CCD is still on there, broken. Would probably help dissipate heat too.
junpit said:
I’ve got a 5800x3d and a 4090. Anyone think I should upgrade to a 7800x3d? I play on 4K. My thoughts are that I don’t need to and the performance improvement is negligible at that resolution.
Click to expand...
You answered your own question. If I was building now, I'd probably get it for the lolz. I watch DF and they make me feel embarrassed for owning a r5 3600, saying "that version of Ryzen was never great for gaming", unlike the praise at launch. But I let games mature rather than buy at release. I know I don't "need" an upgrade, and probably wouldn't feel it day to day. If each Zen provides 15% improvements at best, I'm waiting 5 generations (Zen 7) for a doubling of performance on the x600 chip.
 
Last edited:
Gerard said:
When amd released the 5800x3d people moaned there was no higher core count version. They delivered a higher core count version this gen. What they did wrong was market them as gaming cpus as their obvious strength is in other areas.

I'm also sure they stated they tested vcache on multi ccd cpus internally and the results weren't much better than what they released with the 7950X3D/7900X3D vs the additional cost of adding vcache to the other ccds.

Just have to wait and see what comes next they'll obviously develop it further.
Click to expand...
I think when people wanted higher core counts they were imagining they would have the VCache on all the cores, not 50% of them and then have to turn off 50% of the cores.

Rawr said:
I get the feeling there's some scheduler bugs going on where it's maybe moving it off the main cores or something. That was kind of apparent from the reviews of the 7950x3d where they just disabled the second CCD to see potential 7800 perf. You'd probably have to compare ccd disabled 7950 to 7800 to get an exact out of the box performance comparison for the cached ccd. That all said as I said previously I'm of the opinion minor clock speed differences is probably not going to be that impactful regardless.
Click to expand...
I think LtMatt confirmed earlier in this thread that the scheduler is working perfectly.
 
GoogalyMoogaly said:
I think when people wanted higher core counts they were imagining they would have the VCache on all the cores, not 50% of them and then have to turn off 50% of the cores.


I think LtMatt confirmed earlier in this thread that the scheduler is working perfectly.
Click to expand...
I didn't see all of his posts but a lot of them seemed related to tinkering a bit, not totally out of the box.

I didn't think even the most diehard AMD fans are saying the scheduler is working perfectly. It's definitely working well enough if you don't mind a slight bit of tinkering here and there though sure.
 
Gerard said:
When amd released the 5800x3d people moaned there was no higher core count version. They delivered a higher core count version this gen. What they did wrong was market them as gaming cpus as their obvious strength is in other areas.

I'm also sure they stated they tested vcache on multi ccd cpus internally and the results weren't much better than what they released with the 7950X3D/7900X3D vs the additional cost of adding vcache to the other ccds.

Just have to wait and see what comes next they'll obviously develop it further.
Click to expand...

I don't think that was wrong.

The moment you stack it against the non-3D version you see it's significantly worse in non-gaming tasks so how could they have marketed it as anything else.

As I see it the wrong part is listening to the fans, realising it was stupid to go double cached chiplets but still going ahead by making a weird and expensive half/half product looking for a purpose.

The existence of the 7950X3D has meant the 7800X3D had to get a 200MHz penalty to reduce shaming of the more expensive cpu in gaming performance they're both sold for. That's the cost of forcing this compromise flagship into existence, the better and cheaper product had to be made worse.
 
Narn said:
any test results of 7800x3d with fs2020? is it even better performing than 7950x3d?
Click to expand...
flightsimqyiht.png
 
Hotwired said:
I don't think that was wrong.

The moment you stack it against the non-3D version you see it's significantly worse in non-gaming tasks so how could they have marketed it as anything else.

As I see it the wrong part is listening to the fans, realising it was stupid to go double cached chiplets but still going ahead by making a weird and expensive half/half product looking for a purpose.

The existence of the 7950X3D has meant the 7800X3D had to get a 200MHz penalty to reduce shaming of the more expensive cpu in gaming performance they're both sold for. That's the cost of forcing this compromise flagship into existence, the better and cheaper product had to be made worse.
Click to expand...

Was referring to marketing the 7900x3d/7950x3d as gaming chips on launch, a few other outlets have stated amd should have been trumpeting their efficiency as the headline for them rather than the gaming focus.
 
Rawr said:
I didn't see all of his posts but a lot of them seemed related to tinkering a bit, not totally out of the box.

I didn't think even the most diehard AMD fans are saying the scheduler is working perfectly. It's definitely working well enough if you don't mind a slight bit of tinkering here and there though sure.
Click to expand...
I can't remember the post exactly but I'm sure he said it's work exactly how it's intended. You know how Matt has his overlay so he can see if the cores are sleeping.
 
Hotwired said:
I don't think that was wrong.

The moment you stack it against the non-3D version you see it's significantly worse in non-gaming tasks so how could they have marketed it as anything else.

As I see it the wrong part is listening to the fans, realising it was stupid to go double cached chiplets but still going ahead by making a weird and expensive half/half product looking for a purpose.

The existence of the 7950X3D has meant the 7800X3D had to get a 200MHz penalty to reduce shaming of the more expensive cpu in gaming performance they're both sold for. That's the cost of forcing this compromise flagship into existence, the better and cheaper product had to be made worse.
Click to expand...
More likely binned chips with the higher clock.
 
Gerard said:
Was referring to marketing the 7900x3d/7950x3d as gaming chips on launch, a few other outlets have stated amd should have been trumpeting their efficiency as the headline for them rather than the gaming focus.
Click to expand...

I think that could have brought up awkward comparisons with just turning on eco mode on the regular cpus that already cost less since those came out using excessive power.

Or maybe it was a medium length con and the regular cpus were released using extra power on purpose to create a larger efficiency appearance when the X3D came out using correctly set power limits, lower frequency and 3D cache.
 
GoogalyMoogaly said:
I can't remember the post exactly but I'm sure he said it's work exactly how it's intended. You know how Matt has his overlay so he can see if the cores are sleeping.
Click to expand...
The chips only got released into the wild a month ago. Optimisations probably haven’t even been looked at yet.
 
Last edited:
delta0 said:
More likely binned chips with the higher clock.
Click to expand...

It's only 200MHz lower frequency at 5005 and we've seen that external clock motherboards can overclock the 7800X3D and have reached 5400 last I looked.

I believe a segmentation decision is more likely than being able to find another 400MHz like that.

The main guy demoing his overclocking has got it up to 5600 with no load and probably too high voltage but unless he's got a solid gold cpu it means limiting it to 5000 while the flagship gets a 5200 limit is in my opinion, more artificial than it could be

tmp.png
 
Last edited:
Reviews seem to be a bit grumbly that the 7800X3D is slower than the 7700X in productivity... but a lot of the difference is less than the % of clockspeed drop would imply. And I'm not really sure that a person just using the PC will actually feel that? Like if I do decide to have a go with Blender or Photoshop, it's still blooming fast (especially compared to my 8700k)...

Very impressed by the power figures, seeing like 80W quoted as the max they've seen it draw. That's crackers for what it's delivering. I'd just about settled to go full sized X670 but all of a sudden I'm pondering ITX possibilities with that little power draw...

Hotwired said:
we've seen that external clock motherboards can overclock the 7800X3D and have reached 5400 last I looked.
Click to expand...

Is... is there a shortlist of the boards that can/can't make this happen? I will probably never bother, but it would be nice to know I've bought something that can :D
 
Last edited:
Hotwired said:
It's only 200MHz lower frequency at 5005 and we've seen that external clock motherboards can overclock the 7800X3D and have reached 5400 last I looked.

I believe a segmentation decision is more likely than being able to find another 400MHz like that.

The main guy demoing his overclocking has got it up to 5600 with no load and probably too high voltage but unless he's got a solid gold cpu it means limiting it to 5000 while the flagship gets a 5200 limit is in my opinion, more artificial than it could be

tmp.png
Click to expand...
That could well be right AMD have done that before. The same person that got 5.4 GHz on the 7800X3D got 5.45GHz on the 3D cached CCD0 for the 7950X3D. I’m sure if we dig we would find more with higher numbers. I think it’s a brave move to push these chips too much. Could get expensive very quickly! :eek:
 
You must log in or register to reply here.
Back
Top Bottom