Zen 3

2020 AMD 7-nanometer processor microarchitecture From Wikipedia, the free encyclopedia

Zen 3 is the name for a CPU microarchitecture by AMD, released on November 5, 2020.[2][3] It is the successor to Zen 2 and uses TSMC's 7 nm process for the chiplets and GlobalFoundries's 14 nm process for the I/O die on the server chips and 12 nm for desktop chips.[4] Zen 3 powers Ryzen 5000 mainstream desktop processors (codenamed "Vermeer") and Epyc server processors (codenamed "Milan").[5][6] Zen 3 is supported on motherboards with 500 series chipsets; 400 series boards also saw support on select B450 / X470 motherboards with certain BIOSes.[7] Zen 3 is the last microarchitecture before AMD switched to DDR5 memory and new sockets, which are AM5 for the desktop "Ryzen" chips alongside SP5 and SP6 for the EPYC server platform and sTRX8.[3] According to AMD, Zen 3 has a 19% higher instructions per cycle (IPC) on average than Zen 2.

LaunchedNovember 5, 2020; 5 years ago (November 5, 2020)
Designed byAMD
Common manufacturers
CPUID codeFamily 19h
Quick facts General information, Launched ...
AMD Zen 3
General information
LaunchedNovember 5, 2020; 5 years ago (November 5, 2020)
Designed byAMD
Common manufacturers
CPUID codeFamily 19h
Physical specifications
Transistors
  • 6.24 billion (1× CCD) or
    10.39 billion (2× CCD)
    (4.15 billion per 7 nm 8-core "CCD" & 2.09 billion for the 12 nm "I/O die")[1]
Cores
    • Desktop: 4 to 16
    • Workstation: 16 to 64
    • Server: 16 to 64
Package
  • Package FP6
Sockets
Cache
L1 cache64 KB (per core):
  • 32 KB instructions
  • 32 KB data
L2 cache512 KB (per core)
L3 cache
  • 32 MB (per CCD)
  • 96 MB (per CCD with 3D V-Cache)
  • 16 MB (in APUs)
Architecture and classification
Technology node
MicroarchitectureZen
Instruction setAMD64 (x86-64)
Products, models, variants
Product code names
  • Desktop
    • Vermeer (w/o iGPU)
    • Cézanne (APU)
    • Chagall

  • Thin & Light Mobile
    • Cézanne
    • Barcelo
    • Barcelo-R

  • High-End Mobile
    Cézanne
  • Server
    • Milan
    • Milan-X

History
PredecessorZen 2
Successors
Support status
Supported
Close

On April 1, 2022, AMD released the new Ryzen 6000 series for laptops/mobile, using an improved Zen 3+ architecture featuring notable architectural improvements to power efficiency and power management.[8] And slightly later, on April 20, 2022, AMD would also release the Ryzen 7 5800X3D desktop processor, which increased gaming performance by around +15% on average by using for the very first time in a PC product, a 3D vertically stacked L3 cache. Specifically in the form of a 64MB L3 cache "3D V Cache" die made on the same TSMC N7 process as the 8-core Zen 3 CCD which it gets direct copper to copper hybrid bonded to.[9]

Features

As the first largely "ground up redesign" of the Zen CPU core since the architecture family's original release in early 2017 with Zen 1/Ryzen 1000, Zen 3 was a significant architectural improvement over its predecessors; having a very significant IPC increase of +19% over the prior Zen 2 architecture in addition to being capable of reaching higher clock speeds.[10]

Like Zen 2, Zen 3 is composed of up to 2 core complex dies (CCD) along with a separate IO die containing the I/O components. A Zen 3 CCD is composed of a single core complex (CCX) containing 8 CPU cores and 32 MB of shared L3 cache, this is in contrast to Zen 2 where each CCD is composed of 2 CCX, each containing 4 cores paired with 16 MB of L3 cache. The new configuration allows all 8 cores of the CCX to directly communicate with each other and the L3 Cache instead of having to use the IO die through the Infinity Fabric.[10]

Zen 3 (along with AMD's RDNA2 GPUs) also implemented Resizable BAR, an optional feature introduced in PCIe 2.0, that was branded as Smart Access Memory (SAM). This technology allows CPU to directly access all of compatible video card's VRAM.[11] Intel and Nvidia have since implemented this feature as well.[12]

  • A de-lidded Ryzen 5 5600X. Only one 6-core CCD is present. The contacts for a second CCD are visible.
    A de-lidded Ryzen 5 5600X. Only one 6-core CCD is present. The contacts for a second CCD are visible.
  • Close-up of the CCD, taken under infrared lighting. This die was damaged by the de-lidding process.
    Close-up of the CCD, taken under infrared lighting. This die was damaged by the de-lidding process.
  • Close-up of the I/O die
    Close-up of the I/O die

In Zen 3, a single 32MB L3 cache pool is shared among all 8 cores in a chiplet, vs. Zen 2's two 16MB pools each shared among 4 cores in a core complex, of which there were two per chiplet. This new arrangement improves the cache hit rate as well as performance in situations that require cache data to be exchanged among cores, but increases cache latency from 39 cycles in Zen 2 to 46 clock cycles and halves per-core cache bandwidth, although both problems are partially mitigated by higher clock speeds. Total cache bandwidth on all 8 cores combined remains the same due to power consumption concerns. L2 cache capacity and latency remain the same at 512KB and 12 cycles. All cache read and write operations are done at 32 bytes per cycle.[13]

On April 20, 2022, AMD released the R7 5800X3D. It features, for the first time in a desktop PC product, 3D-stacked vertical L3 cache. Its extra 64MB comes via a TSMC N7 (7nm) "3D V Cache" die direct copper to copper hybrid bonded right on top of the 8-core Zen 3 CCD's usual 32MB, increasing the CPU's total L3 cache capacity to 96MB and bringing significant performance improvements for gaming in particular; now rivalling contemporary high-end consumer processors while being much more power efficient and running on older, cheaper motherboards using affordable DDR4 memory.[9] And despite now spanning multiple dies and being three times larger (96MB vs 32MB), the L3 cache's performance remains nearly identical; with X3D only adding around ≈+2ns via an additional three to four cycles of latency.[14] It would later be followed by the Ryzen 5 5600X3D and Ryzen 7 5700X3D for lower-end market segments, and succeeded by the Ryzen 7000X3D family of 3D V Cache equipped Zen 4 processors on the newer socket AM5 platform.

Improvements

CCD layouts comparison for Zen 2 and Zen 3

Zen 3 has made the following improvements over Zen 2:[13][15]

  • An increase of 19% in instructions per clock
  • The base core chiplet has a single eight-core complex (versus two four-core complexes in Zen 2)
  • A unified 32MB L3 cache pool equally available to all 8 cores in a chiplet, vs Zen 2's two 16MB pools each shared among 4 cores in a core complex.
    • On mobile: A unified 16MB L3
  • A unified 8-core CCX (from 2x 4-core CCX per CCD)
  • Increased branch prediction bandwidth. L1 branch target buffer size increased to 1024 entries (vs 512 in Zen 2)
  • New instructions
  • Improved integer units
    • 96 entry integer scheduler (up from 92)
    • 192 entry physical register file (up from 180)
    • 10 issue per cycle (up from 7)
    • 256 entry reorder-buffer (up from 224)
    • fewer cycles for DIV/IDIV ops (10...20 from 16...46)
  • Improved floating point units
    • 6 μOP dispatch width (up from 4)
    • FMA latency reduced by 1 cycle (down from 5 to 4)
  • Additional 64MB 3D vertically stacked dense library L3 cache (in -X3D models)

Feature tables

CPUs

APUs

APU features table

Products

AMD Ryzen 7 5800X

On October 8, 2020, AMD announced four Zen 3-based desktop Ryzen processors, consisting of one Ryzen 5, one Ryzen 7, and two Ryzen 9 CPUs and featuring between 6 and 16 cores.[2]

Desktop CPUs

Vermeer

Common features of Ryzen 5000 desktop CPUs:

  • Socket: AM4.
  • All the CPUs support DDR4-3200 in dual-channel mode.
  • All the CPUs support 24 PCIe 4.0 lanes. 4 of the lanes are reserved as link to the chipset.
  • No integrated graphics.
  • L1 cache: 64 KB per core (32 KB data + 32 KB instruction).
  • L2 cache: 512 KB per core.
  • Fabrication process: TSMC 7FF.
More information Branding and model, Cores (threads) ...
Branding and model Cores
(threads)		 Clock rate (GHz) L3 cache
(total)		 TDP Chiplets Core
config[i]		 Thermal
solution		 Release
date		 MSRP
Base Boost
Ryzen 9 5950X 16 (32) 3.4 4.9 64 MB 105 W 2 × CCD
1 × I/OD		 2 × 8 N/a Nov 5, 2020 US $799
5900XT 3.3 4.8 Jul 31, 2024 US $349
5900X 12 (24) 3.7 2 × 6 Nov 5, 2020 US $549
5900 3.0 4.7 65 W Jan 12, 2021 OEM
PRO 5945 Sep 2022[16]
Ryzen 7 5800X3D 8 (16) 3.4 4.5 96 MB 105 W 1 × CCD
1 × I/OD		 1 × 8 Apr 20, 2022 US $449
5800XT 3.8 4.8 32 MB Wraith Prism, None[ii] Jul 31, 2024 US $249
5800X 4.7 N/a Nov 5, 2020 US $449
5800 3.4 4.6 65 W Jan 12, 2021 OEM
5700X3D 3.0 4.1 96 MB 105 W Jan 31, 2024[18] US $249
5700X 3.4 4.6 32 MB 65 W Apr 4, 2022 US $299
PRO 5845 Sep 2022 OEM
Ryzen 5 5600X3D 6 (12) 3.3 4.4 96 MB 105 W 1 × 6 Jul 7, 2023 US $229[19]
US Only[20]
5600XT 3.7 4.7 32 MB 65 W Wraith Stealth[21] Oct 31, 2024 US $194[22]
5600X 3.7 4.6 Nov 5, 2020 US $299
5600T 3.5 4.5 Oct 31, 2024 US $186[22]
5600 3.5 4.4 Apr 4, 2022 US $199
5600F 3.0 4.0 Sep 16, 2025 APJ Only[23]
PRO 5645 3.7 4.6 N/a Sep 2022 OEM
5500X3D 3.0 4.0 96 MB 105 W Jun 5, 2025 LATAM Only[24]
Close
  1. [i]
    Core Complexes (CCX) × cores per CCX
  2. [ii]
    On August 1, 2025, AMD downgraded or ceased including coolers on select models, as a result of its discontinuation of the Wraith Prism and Wraith Spire.[17]

Cezanne

Based on the Ryzen 5000 series APUs but with the integrated graphics disabled. Common features of Ryzen 5000 (Cezanne) desktop CPUs:

  • Socket: AM4.
  • CPUs support DDR4-3200 in dual-channel mode.
  • L1 cache: 64 KB (32 KB data + 32 KB instruction) per core.
  • L2 cache: 512 KB per core.
  • CPUs support 24 PCIe 3.0 lanes. 4 of the lanes are reserved as link to the chipset.
  • No integrated graphics.
  • Fabrication process: TSMC 7FF.
More information Branding and model, Cores (threads) ...
Branding and model Cores
(threads)		 Thermal
solution		 Clock rate (GHz) L3 cache
(total)		 TDP Core
config[i]		 Release
date		 MSRP
(USD)
Base Boost
Ryzen 7 5700[25] 8 (16) Wraith Stealth 3.7 4.6 16 MB 65 W 1 × 8 Apr 4, 2022 (OEM),
Dec 21, 2023 (retail)		 $179
Ryzen 5 5500 6 (12) 3.6 4.2 1 × 6 Apr 4, 2022 $159
Ryzen 3 5100[26][27][28] 4 (8) – 3.8 8 MB 1 × 4 2023 OEM
Close
  1. [i]
    Core Complexes (CCX) × cores per CCX

5100, 5500, and 5700 have no ECC support like non-Pro Ryzen 5000 Desktop APUs.

Chagall

Common features of Ryzen 5000 workstation CPUs:

  • Socket: sWRX8.
  • All the CPUs support DDR4-3200 in octa-channel mode.
  • L1 cache: 64 KB (32 KB data + 32 KB instruction) per core.
  • L2 cache: 512 KB per core.
  • All the CPUs support 128 PCIe 4.0 lanes. 8 of the lanes are reserved as link to the chipset.
  • No integrated graphics.
  • Fabrication process: TSMC 7FF.
More information Branding and Model, Cores (threads) ...
Branding and Model Cores
(threads)		 Clock rate (GHz) L3 cache
(total)		 TDP Chiplets Core
config[i]		 Release
date		 MSRP
Base Boost
Ryzen
Threadripper
PRO		 5995WX 64 (128) 2.7 4.5 256 MB 280 W 8 × CCD
1 × I/OD		 8 × 8 Mar 8, 2022
(OEM) /
?
(retail)		 OEM /
US $6500
5975WX 32 (64) 3.6 128 MB 4 × CCD
1 × I/OD		 4 × 8 Mar 8, 2022
(OEM) /
?
(retail)		 OEM /
US $3300
5965WX 24 (48) 3.8 4 × 6 Mar 8, 2022
(OEM) /
?
(retail)		 OEM /
US $2400
5955WX 16 (32) 4.0 64 MB 2 × CCD
1 × I/OD		 2 × 8 Mar 8, 2022 OEM
5945WX 12 (24) 4.1 2 × 6
Close
  1. [i]
    Core Complexes (CCX) × cores per CCX

Desktop APUs

Cezanne

Common features of Ryzen 5000 desktop APUs:

  • Socket: AM4.
  • All the CPUs support DDR4-3200 in dual-channel mode.
  • L1 cache: 64 KB (32 KB data + 32 KB instruction) per core.
  • L2 cache: 512 KB per core.
  • All the CPUs support 24 PCIe 3.0 lanes. 4 of the lanes are reserved as link to the chipset.
  • Includes integrated GCN 5th generation GPU.
  • Fabrication process: TSMC 7FF.
More information Branding and model, CPU ...
Branding and model CPU GPU[a] Thermal
solution		 TDP Release
date		 MSRP
Cores
(threads)		 Clock rate (GHz) L3 cache
(total)		 Core
config[i]		 Clock
(MHz)		 Config[ii] Processing
power[iii]
(GFLOPS)
Base Boost
Ryzen 7 5705G 8 (16) 3.8 4.6 16 MB 1 × 8 2000 512:32:8
8 CU		 2048 N/a 65 W
5700G[b] Wraith Stealth Apr 13, 2021 (OEM),
Aug 5, 2021 (retail)		 US $359
5705GE 3.2 N/a 35 W
5700GE[b] Wraith Stealth Apr 13, 2021 OEM
Ryzen 5 5600GT 6 (12) 3.6 1 × 6 1900 448:28:8
7 CU		 1702.4 65 W Jan 31, 2024[29] US $140
5605G 3.9 4.4 N/a
5600G[b] Wraith Stealth Apr 13, 2021 (OEM),
Aug 5, 2021 (retail)		 US $259
5605GE 3.4 N/a 35 W
5600GE[b] Wraith Stealth Apr 13, 2021 OEM
5500GT 3.6 65 W Jan 31, 2024[29] US $125
Ryzen 3 5305G 4 (8) 4.0 4.2 8 MB 1 × 4 1700 384:24:8
6 CU		 1305.6 N/a
5300G[b] OEM Apr 13, 2021 OEM
5305GE 3.6 N/a 35 W
5300GE[b] OEM Apr 13, 2021 OEM
Close
  1. [i]
    Core Complexes (CCX) × cores per CCX
  2. [ii]
    Unified shaders : texture mapping units : render output units and compute units (CU)
  3. [iii]
    Single precision performance is calculated from the base (or boost) core clock speed based on a FMA operation.
  1. [a]
    All of the iGPUs are branded as AMD Radeon Graphics.
  2. [b]
    Model also available as PRO version as 5350GE,[30] 5350G,[31] 5650GE,[32] 5650G,[33] 5750GE,[34] 5750G,[35] released June 1, 2021. Refresh models 5355GE, 5355G, 5655GE, 5655G, 5755GE, 5755G released September 5, 2024.[36]

Mobile APUs

Cezanne

More information Branding and Model, CPU ...
Branding and Model CPU GPU TDP Release
date
Cores
(Threads)		 Clock rate (GHz) L3 cache
(total)		 Core
config[i]		 Model Clock
(GHz)		 Config[ii] Processing
power
(GFLOPS)[iii]
Base Boost
Ryzen 9 5980HX[37] 8 (16) 3.3 4.8 16 MB 1 × 8 Radeon
Graphics
[a]		 2.1 512:32:8
8 CUs		 2150.4 35–54 W Jan 12, 2021
5980HS[38] 3.0 35 W
5900HX[39] 3.3 4.6 35–54 W
5900HS[40] 3.0 35 W
Ryzen 7 5800H[41][42] 3.2 4.4 2.0 2048 35–54 W
5800HS[43] 2.8 35 W
5800U[note 1][44] 1.9 10–25 W
Ryzen 5 5600H[45][46] 6 (12) 3.3 4.2 1 × 6 1.8 448:28:8
7 CUs		 1612.8 35–54 W
5600HS[47] 3.0 35 W
5600U[note 1][48] 2.3 10–25 W
5560U[49] 4.0 8 MB 1.6 384:24:8
6 CUs		 1228.8
Ryzen 3 5400U[note 1][50][51] 4 (8) 2.7 4.1 1 × 4
Close
  1. [a]
    All of the iGPUs are branded as AMD Radeon Graphics.
  1. [i]
    Core Complexes (CCX) × cores per CCX
  2. [ii]
    Unified shaders : texture mapping units : render output units and compute units (CU)
  3. [iii]
    Single precision performance is calculated from the base (or boost) core clock speed based on a FMA operation.
  1. [1]
    Model also available as PRO version as 5450U,[52] 5650U,[53] 5850U,[54] released on March 16, 2021.

Barceló

More information Branding and model, CPU ...
Branding and model CPU GPU TDP Release
date
Cores
(Threads)		 Clock rate (GHz) L3 cache
(total)		 Core
config[i]		 Model Clock
(GHz)		 Config[ii] Processing
power
(GFLOPS)[iii]
Base Boost
Ryzen 7 5825U[note 1][note 2][55] 8 (16) 2.0 4.5 16 MB 1 × 8 Radeon
Graphics[a]		 2.0 512:32:8
8 CUs		 2048 15 W Jan 4, 2022
Ryzen 5 5625U[note 1][note 2][56] 6 (12) 2.3 4.3 1 × 6 1.8 448:28:8
7 CUs		 1612.8
Ryzen 3 5425U[57] 4 (8) 2.7 4.1 8 MB 1 × 4 1.6 384:24:6
6 CUs		 ? Jan 30, 2022
Ryzen 3 5125C[58] 2 (4) 3.0 N/a 1 × 2 ? 192:12:8
3 CU		 ? May 5, 2022
Close
  1. [a]
    All of the iGPUs are branded as AMD Radeon Graphics.
  1. [i]
    Core Complexes (CCX) × cores per CCX
  2. [ii]
    Unified shaders : texture mapping units : render output units and compute units (CU)
  3. [iii]
    Single precision performance is calculated from the base (or boost) core clock speed based on a FMA operation.
  1. [1]
    Model also available as Pro version as 5475U,[59] 5675U,[60] 5875U,[61] released on April 19, 2022.
  2. [2]
    Model also available as Chromebook optimized version as 5425C,[62] 5625C,[63] 5825C,[64] released on May 5, 2022.

Barceló-R

Common features of Ryzen 7030 notebook APUs:

  • Socket: FP6.
  • All the CPUs support DDR4-3200 or LPDDR4-4266 in dual-channel mode.
  • L1 cache: 64 KB (32 KB data + 32 KB instruction) per core.
  • L2 cache: 512 KB per core.
  • All the CPUs support 16 PCIe 3.0 lanes.
  • Native USB 4 (40Gbps) Ports: 0
  • Native USB 3.2 Gen 2 (10Gbps) Ports: 2
  • Includes integrated GCN 5th generation GPU.
  • Fabrication process: TSMC N7 FinFET.
More information Branding and Model, CPU ...
Branding and Model CPU GPU TDP Release
date
Cores
(threads)		 Clock rate (GHz) L3 cache
(total)		 Core
config[a]		 Model Clock
(GHz)		 Processing
power[b]
(GFLOPS)
Base Boost
Ryzen 7 (PRO) 7730U 8 (16) 2.0 4.5 16 MB 1 × 8 Vega
8 CU		 2.0 2048 15 W January 4, 2023
[65]
Ryzen 5 (PRO) 7530U 6 (12) 1 × 6 Vega
7 CU		 1792
7430U 2.3 4.3 1.8 1612.8 Q4 2023
Ryzen 3 (PRO) 7330U 4 (8) 8 MB 1 × 4 Vega
6 CU		 1382.4 January 4, 2023
[66]
Close
  1. [a]
    Core Complexes (CCX) × cores per CCX
  2. [b]
    Single precision performance is calculated from the base (or boost) core clock speed based on a FMA operation.

Embedded CPUs

More information Model, Release date ...
Model Release
date		 Fab CPU Socket PCIe
support		 Memory
support		 TDP
Cores
(threads)		 Clock rate (GHz) Cache
Base Boost L1 L2 L3
V3C14[67][68] September 27, 2022[69] TSMC
7FF		 4 (8) 2.3 3.8 32 KB inst.
32 KB data
per core		 512 KB
per core		 8 MB FP7r2 20
(8+4+4+4)
PCIe 4.0		 DDR5-4800
dual-channel		 15 W
V3C44[67][68] 3.5 3.8 45 W
V3C16[67][68] 6 (12) 2.0 3.8 16 MB 15 W
V3C18I[67][68] 8 (16) 1.9 3.8 15 W
V3C48[67][68] 3.3 3.8 45 W
Close

    Server CPUs

    The Epyc server line of chips based on Zen 3 is named Milan and is the final generation of chips using the SP3 socket.[6] Epyc Milan was released on March 15, 2021.[70]

    Common features:

    • SP3 socket
    • Zen 3 microarchitecture
    • TSMC 7 nm process for the compute and cache dies, GloFo 14 nm process for the I/O die
    • MCM with one I/O Die (IOD) and multiple Core Complex Dies (CCD) for compute, one core complex (CCX) per CCD chiplet
    • Eight-channel DDR4-3200
    • 128 PCIe 4.0 lanes per socket, 64 of which are used for Infinity Fabric inter-processor links in 2P platforms
    • 7003X series models include 64 MiB L3 cache dies stacked on top of the compute dies (3D V-Cache)
    • 7003P series models are limited to uniprocessor operation (1P, single-socket)
    More information Model, Cores (threads) ...
    Model Cores
    (threads)    		 Chiplets Core
    config
    [i]    		 Clock rate Cache size Socket Scaling TDP
    default (range)    		 Release
    price
    Base
    (GHz)    		 Boost
    (GHz)    		 L2
    per core    		 L3
    per CCX    		 Total
    7203(P)   8 (16)2 + IOD 2 × 4  2.8 3.4 512 KB32 MB68 MB  SP32P (1P) 120 W (120-150)  $348 ($338)
    72F3 8 + IOD 8 × 1  3.7 4.1260 MB2P 180 W (165-200)$2468
    7303(P) 16 (32)2 + IOD 2 × 8  2.4 3.432 MB72 MB2P (1P) 130 W (120-150)  $604 ($594)
    7313(P) 4 + IOD 4 × 4  3.0 3.7136 MB2P (1P) 155 W (155-180)$1083 ($913)
    7343  3.2 3.92P 190 W (165-200)$1565
    73F3 8 + IOD 8 × 2  3.5 4.0264 MB 240 W (225-240)$3521
    7373X 8* + IOD  3.05 3.896 MB776 MB 240 W (225-280)$4185
    7413 24 (48)4 + IOD 4 × 6  2.65 3.632 MB140 MB2P 180 W (165-200)$1825
    7443(P)  2.85 4.02P (1P) 200 W (165-200)$2010 ($1337)
    74F3 8 + IOD 8 × 3  3.2 4.0268 MB2P 240 W (225-240)$2900
    7473X 8* + IOD  2.8 3.796 MB780 MB 240 W (225-280)$3900
    7453 28 (56)4 + IOD 4 × 7  2.75 3.4516 MB78 MB2P 225 W (225-240)$1570
    7513 32 (64)4 + IOD 4 × 8  2.6 3.6532 MB144 MB2P200 W (165-200)$2840
    7543(P) 8 + IOD 8 × 4  2.8 3.7272 MB2P (1P) 225 W (225-240)$3761 ($2730)
    75F3  2.95 4.02P 280 W (225-280)$4860
    7573X 8* + IOD  2.8 3.696 MB784 MB $5590
    7R13[71] 48 (96)6 + IOD 6 × 8  2.65 3.732 MB216 MBTBD TBDOEM/AWS
    7643(P) 8 + IOD 8 × 6  2.3 3.6280 MB2P (1P) 225 W (225-240)$4995 ($2722)
    7663 56 (112)8 + IOD 8 × 7  2.0 3.532 MB284 MB2P 240 W (225-240)$6366
    7663P 1P240 W (225-280)$3139
    7713(P) 64 (128)8 + IOD 8 × 8  2.0 3.67532 MB288 MB2P (1P) 225 W (225-240)$7060 ($5010)
    7763  2.45 3.42P 280 W (225-280)$7890
    7773X 8* + IOD  2.2 3.596 MB800 MB $8800
    Close
    1. [i]
      Core Complexes (CCX) × cores per CCX

    Zen 3+

    Quick facts General information, Launched ...
    AMD Zen 3+
    General information
    LaunchedApril 1, 2022; 3 years ago (April 1, 2022)
    Designed byAMD
    Common manufacturer
    CPUID codeFamily 19h
    Physical specifications
    Cores
    • 4 to 8
    Package
      • Package FP7
      • Package FP7r2
    Cache
    L1 cache64 KB (per core)
    L2 cache512 KB (per core)
    L3 cacheUp to 16 MB
    Architecture and classification
    Technology nodeTSMC N6
    Instruction setAMD64 (x86-64)
    Products, models, variants
    Product code names
    • Thin & Light Mobile
      • Rembrandt
      • Rembrandt-R

    • High-End Mobile
      • Rembrandt
      • Rembrandt-R

    History
    PredecessorZen 3
    SuccessorZen 4
    Support status
    Supported
    Close

    Zen 3+ is the codename for a refresh of the Zen 3 microarchitecture, which focuses on power efficiency improvements. It was released in April 2022 with the Ryzen 6000 series of mobile processors.

    Features and improvements

    Zen 3+ has 50 new or enhanced power management features over Zen 3, and also provides an adaptive power management framework, as well as new deep sleep states. Altogether, this brings improvements to efficiency both during idle, and when under load, with up to 30% performance-per-watt increase over Zen 3, as well as longer battery life.[72][73]

    IPC is identical to that of Zen 3; the performance improvements of Ryzen 6000 over Ryzen 5000 mobile processors stem from it having a higher efficiency (hence more performance in power-constrained form factors like laptops), as well as the increased clock speeds from being built on the smaller TSMC N6 node.[74]

    The Rembrandt implementation of Zen 3+ also has support for DDR5 and LPDDR5 memory.

    iGPU of Zen 3+ can support AV1 hardware decoding.

    Products

    Warhol (Cancelled)

    In mid-2020, reports indicated that AMD was preparing a refreshed lineup of Zen 3-based desktop AM4 processors, codenamed “Warhol.” This family was widely expected to be launched as the Ryzen 6000 series for desktops and to serve as the direct successor to “Vermeer,” the original Zen 3 Ryzen CPUs. According to a leaked internal roadmap, Zen 3+ (Warhol) appeared to position as an intermediate step between Zen 3 (Vermeer) and the next major architecture, Zen 4 (Raphael).[75]

    Warhol was described as a modest architectural refresh rather than a full redesign with main enhancements centered on transitioning from TSMC’s 7 nm process, to the more efficient TSMC 6 nm (N6) node. While N6 does not introduce major architectural changes, it offers better transistor density and improved power characteristics. As a result, Warhol series was expected to deliver incremental gains in performance—similar in scale to AMD’s earlier Zen+ refresh (Pinnacle Ridge, released as the Ryzen 2000 series), which provided refined performance without altering the core microarchitecture.[75][76]

    However it was reported that AMD ultimately cancelled the Warhol lineup. Multiple factors likely contributed to this decision. The 2020–2023 global chip shortage placed immense strain on foundry capacity worldwide, including TSMC’s ability to satisfy the demand for its 7 nm and 6 nm nodes. AMD decided to focus its resources toward the next major architectural leap: Zen 4 (Raphael), built on TSMC’s 5 nm process. This effectively ended the development of Warhol in favor of the arrival of AMD’s next major CPU lineup.[77][78]

    Rembrandt

    On April 1, 2022, AMD released the Ryzen 6000 series of mobile APUs, codenamed Rembrandt. It introduces PCIe 4.0 and DDR5/LPDDR5 for the first time in an APU for the laptop and also introduced RDNA2 integrated graphics to the PC. It is built on TSMC's 6 nm node.[8]

    Common features of Ryzen 6000 notebook APUs:

    • Socket: FP7, FP7r2.
    • All the CPUs support DDR5-4800 or LPDDR5-6400 in dual-channel mode.
    • L1 cache: 64 KB (32 KB data + 32 KB instruction) per core.
    • L2 cache: 512 KB per core.
    • All the CPUs support 16 PCIe 4.0 lanes.
    • Native USB 4 (40Gbps) Ports: 2
    • Native USB 3.2 Gen 2 (10Gbps) Ports: 2
    • Includes integrated RDNA 2 GPU.
    • Fabrication process: TSMC N6 FinFET.
    More information Branding and model, CPU ...
    Branding and model CPU GPU TDP Release
    date
    Cores
    (threads)    		 Clock (GHz) L3 cache
    (total)    		 Core
    config[i]    		 Model Clock
    (GHz)    		 Config[ii] Processing
    power
    (GFLOPS)[iii]
    Base Boost
    Ryzen 9 6980HX 8 (16) 3.3 5.0 16 MB 1 × 8 680M 2.4 768:48:8
    12 CUs    		 3686.4 45 W Jan 4, 2022
    [79]
    6980HS 35 W
    6900HX[a] 4.9 45 W
    6900HS[a] 35 W
    Ryzen 7 6800H[a] 3.2 4.7 2.2 3379.2 45 W
    6800HS[a] 35 W
    6800U[a] 2.7 15–28 W
    Ryzen 5 6600H[a] 6 (12) 3.3 4.5 1 × 6 660M 1.9 384:24:8
    6 CUs    		 1459.2 45 W
    6600HS[a] 35 W
    6600U[a] 2.9 15–28 W
    Close
    1. [i]
      Core Complexes (CCX) × cores per CCX
    2. [ii]
      Unified shaders : texture mapping units : render output units and compute units (CU)
    3. [iii]
      Single precision performance is calculated from the base (or boost) core clock speed based on a FMA operation.
    1. [a]
      Model also available as PRO version (6650U[80], 6650H[81], 6650HS[82], 6850U[83], 6850H[84], 6850HS[85], 6950H[86], 6950HS[87]), released on April 19, 2022.

    Rembrandt-R

    Rembrandt-R is the codename for a refresh of Rembrandt codenamed processors, released as the Ryzen 7035 series of mobile APUs in January 2023.

    Common features of Ryzen 7035 notebook APUs:

    • Socket: FP7, FP7r2.
    • All the CPUs support DDR5-4800 or LPDDR5-6400 in dual-channel mode.
    • L1 cache: 64 KB (32 KB data + 32 KB instruction) per core.
    • L2 cache: 512 KB per core.
    • All the CPUs support 16 PCIe 4.0 lanes.
    • Native USB 4 (40Gbps) Ports: 0
    • Native USB 3.2 Gen 2 (10Gbps) Ports: 2
    • Includes integrated RDNA 2 GPU.
    • Fabrication process: TSMC N6 FinFET.
    More information Branding and model, CPU ...
    Branding and model CPU GPU TDP Release
    date[88]
    Cores
    (threads)    		 Clock (GHz) L3 cache
    (total)    		 Core
    config[a]    		 Model Clock
    (GHz)    		 Processing
    power[b]
    (GFLOPS)
    Base Boost
    Ryzen 7 7735HS 8 (16) 3.2 4.75 16 MB 1 × 8 680M
    12 CU    		 2.2 3379.2 35–54 W April 30, 2023
    7735H
    7736U 2.7 4.7 15–28 W January 4, 2023
    7735U 4.75 15–30 W
    7435HS 3.1 4.5 N/a 35–54 W 2024[89]
    7435H
    Ryzen 5 7535HS 6 (12) 3.3 4.55 1 × 6 660M
    6 CU    		 1.9 1459.2 April 30, 2023
    7535H
    7535U 2.9 15–30 W January 4, 2023
    7235HS 4 (8) 3.2 4.2 8 MB 1 × 4 N/a 35–53 W 2024[90]
    7235H
    Ryzen 3 7335U 3.0 4.3 660M
    4 CU    		 1.8 921.6 15–30 W January 4, 2023
    Close
    1. [a]
      Core Complexes (CCX) × cores per CCX
    2. [b]
      Single precision performance is calculated from the base (or boost) core clock speed based on a FMA operation.

    References

    Related Articles

    Wikiwand AI