15h CEST today.
Some reviews have leaked already.
There is a rumor that AGESA 1.0.0.3 is needed to unlock full potential of Ryzen 3000 on older boards. While X570 will get it at launch, older motherboards will maybe not immediately. Well see, my board [B350 Strix Gaming F] has got two BIOS updates in the last few weeks.
Some reviews have leaked already.
There is a rumor that AGESA 1.0.0.3 is needed to unlock full potential of Ryzen 3000 on older boards. While X570 will get it at launch, older motherboards will maybe not immediately. Well see, my board [B350 Strix Gaming F] has got two BIOS updates in the last few weeks.
Performance summary chart from ChipHell:
The AIDA memtest resuts confirm the IF upstream link is halved (1/2 bus width or clock throttle). Ryzen 3900X (dual chiplet SKU) doubles the memory write throughput, as expected, compared to the 3700X model and maxes out the dual-channel memory throughput.
The AIDA memtest resuts confirm the IF upstream link is halved (1/2 bus width or clock throttle). Ryzen 3900X (dual chiplet SKU) doubles the memory write throughput, as expected, compared to the 3700X model and maxes out the dual-channel memory throughput.
Last edited:
hoom
Veteran
So storming away increasing the multi-threading lead & caught up a lot/competitive on FPU & single-thread
Less of a memory latency hit than I expected showing up in these tests.
From what I've read so far not seeming like a lot of OC potential, whats the story with Intel OC these days?
Rootax
Veteran
Intel oc is good if you can can cool them.
Seems that Zen2 is not hurt that much by the slow write speed, I wonder how...
And, "auto" seems faster than all-core OC in some reviews. Look like AMD really did a good job with their boost solution.
Really curious about TR version.
Edit : did some of you find some infos about the impact of avx2/256 on performances and heat output ?
Seems that Zen2 is not hurt that much by the slow write speed, I wonder how...
And, "auto" seems faster than all-core OC in some reviews. Look like AMD really did a good job with their boost solution.
Really curious about TR version.
Edit : did some of you find some infos about the impact of avx2/256 on performances and heat output ?
hoom
Veteran
Could this be due to the core now doing 2* loads, 1* store per clock?
I'd like to think that this graph sums up all the reviews perfectly:
IPC performance comparison:
The benchmark is multi-threaded, so the concurrent access requests will saturate the interface regardless of the load/store architecture.
Last edited:
Silent_Buddha
Legend
Hmmm, how does AMD do AGESA versions?
https://www.asrock.com/mb/AMD/X370 Taichi/#BIOS
Looking at my current MB it goes from 1.0.0.6 to 0.0.7.2 to 1.0.0.1. What? Why is the AGESA level lower now (1.0.0.1) than it was at the start of this year (1.0.0.6)? I'm confused. But at least it's nice to know that if I want, I can run the new processors in my board.
Regards,
SB
Rootax
Veteran
From the Anandtech review :
* Due to two different variations of the X570 chipset, one with a 15 W and another with an 11 W TDP, the extra power allows for more PCIe lanes, thus better GPU support overall. One example is the ASUS Pro WS X570-Ace model.
And indeed, on this motherboard, it seems that a x16 for the gpu, and a x8 for another device could work, with the later linked to the chipset :
AMD X570 chipset
1 x PCIe 4.0 x16 (x8 mode) instead of x4 for most motherboard.
Very interesting for me...
EDIT : I'm not the only one in this situation it seems : "Exactly. Unless it is stupidly overpriced, or lacks good power distribution and overclocking features, I plan on getting one of these and a Ryzen 3850x. I plan on using the first slot for my GPU in x16, leaving the second slot empty, and sticking my dual port 10gig fiber Intel x520 NIC in the 8x chipset port."
* Due to two different variations of the X570 chipset, one with a 15 W and another with an 11 W TDP, the extra power allows for more PCIe lanes, thus better GPU support overall. One example is the ASUS Pro WS X570-Ace model.
And indeed, on this motherboard, it seems that a x16 for the gpu, and a x8 for another device could work, with the later linked to the chipset :
AMD X570 chipset
1 x PCIe 4.0 x16 (x8 mode) instead of x4 for most motherboard.
Very interesting for me...
EDIT : I'm not the only one in this situation it seems : "Exactly. Unless it is stupidly overpriced, or lacks good power distribution and overclocking features, I plan on getting one of these and a Ryzen 3850x. I plan on using the first slot for my GPU in x16, leaving the second slot empty, and sticking my dual port 10gig fiber Intel x520 NIC in the 8x chipset port."
Last edited:
The Ryzen 9 3900X is already down to $420 in Denmark.
@Ryan Smith How are the revised benchmarks coming along with the new bios? Anything that stands out?
Update: Just saw the revised benchmarks were posted for the 3900X.
@Ryan Smith How are the revised benchmarks coming along with the new bios? Anything that stands out?
Update: Just saw the revised benchmarks were posted for the 3900X.
Last edited:
I can't make any sense of that. There is hardly any change to memory bandwidths. The read bandwidths are limited by the DRAM, but the write bandwidths should show the impact of the 2:1 ratio clearly.
Either both tests run at 1:1 or both run at 2:1.
Cheers
Similar threads
