diff --git a/sys-utils/lscpu-topology.c b/sys-utils/lscpu-topology.c
index 2e1a3b987..92f10b518 100644
--- a/sys-utils/lscpu-topology.c
+++ b/sys-utils/lscpu-topology.c
@@ -209,13 +209,19 @@ static int cputype_read_topology(struct lscpu_cxt *cxt, struct lscpu_cputype *ct
 }
 
 /* count size of all instancess of the "name" */
-size_t lscpu_get_cache_full_size(struct lscpu_cxt *cxt, const char *name)
+size_t lscpu_get_cache_full_size(struct lscpu_cxt *cxt, const char *name, int *instances)
 {
 	size_t i, sz = 0;
 
+	if (instances)
+		*instances = 0;
+
 	for (i = 0; i < cxt->ncaches; i++) {
-		if (strcmp(cxt->caches[i].name, name) == 0)
+		if (strcmp(cxt->caches[i].name, name) == 0) {
 			sz += cxt->caches[i].size;
+			if (instances)
+				(*instances)++;
+		}
 	}
 
 	return sz;
@@ -539,10 +545,10 @@ int lscpu_read_topology(struct lscpu_cxt *cxt)
 	}
 
 	lscpu_sort_caches(cxt->caches, cxt->ncaches);
-	DBG(GATHER, ul_debugobj(cxt, " L1d: %zu", lscpu_get_cache_full_size(cxt, "L1d")));
-	DBG(GATHER, ul_debugobj(cxt, " L1i: %zu", lscpu_get_cache_full_size(cxt, "L1i")));
-	DBG(GATHER, ul_debugobj(cxt, " L2: %zu", lscpu_get_cache_full_size(cxt, "L2")));
-	DBG(GATHER, ul_debugobj(cxt, " L3: %zu", lscpu_get_cache_full_size(cxt, "L3")));
+	DBG(GATHER, ul_debugobj(cxt, " L1d: %zu", lscpu_get_cache_full_size(cxt, "L1d", NULL)));
+	DBG(GATHER, ul_debugobj(cxt, " L1i: %zu", lscpu_get_cache_full_size(cxt, "L1i", NULL)));
+	DBG(GATHER, ul_debugobj(cxt, " L2: %zu", lscpu_get_cache_full_size(cxt, "L2", NULL)));
+	DBG(GATHER, ul_debugobj(cxt, " L3: %zu", lscpu_get_cache_full_size(cxt, "L3", NULL)));
 
 	return rc;
 }
diff --git a/sys-utils/lscpu.c b/sys-utils/lscpu.c
index cd72a3943..b029e0fc0 100644
--- a/sys-utils/lscpu.c
+++ b/sys-utils/lscpu.c
@@ -511,7 +511,7 @@ static void caches_add_line(struct lscpu_cxt *cxt,
 		{
 			uint64_t sz = 0;
 			if (ca->name)
-				sz = lscpu_get_cache_full_size(cxt, ca->name);
+				sz = lscpu_get_cache_full_size(cxt, ca->name, NULL);
 			if (!sz)
 				break;
 			if (cxt->bytes)
@@ -1045,21 +1045,28 @@ static void print_summary(struct lscpu_cxt *cxt)
 		for (i = 0; i < cxt->ncaches; i++) {
 			const char *name = cxt->caches[i].name;
 			uint64_t sz;
+			int n = 0;
 
 			if (last && strcmp(last, name) == 0)
 				continue;
-			sz = lscpu_get_cache_full_size(cxt, name);
+			sz = lscpu_get_cache_full_size(cxt, name, &n);
 			if (!sz)
 				continue;
 			snprintf(field, sizeof(field), is_term ? _("%s:") : _("%s cache:"), name);
 			if (cxt->bytes)
-				add_summary_x(tb, sec, field, "%" PRIu64, sz);
+				add_summary_sprint(tb, sec, field,
+						P_("%" PRIu64 " (%d instance)",
+						   "%" PRIu64 " (%d instances)", n),
+						sz, n);
 			else {
 				char *tmp = size_to_human_string(
 						SIZE_SUFFIX_3LETTER |
 						SIZE_SUFFIX_SPACE,
 						sz);
-				add_summary_s(tb, sec, field, tmp);
+				add_summary_sprint(tb, sec, field,
+						P_("%s (%d instance)",
+						   "%s (%d instances)", n),
+						tmp, n);
 				free(tmp);
 			}
 			last = name;
diff --git a/sys-utils/lscpu.h b/sys-utils/lscpu.h
index 623c07998..e52c77c22 100644
--- a/sys-utils/lscpu.h
+++ b/sys-utils/lscpu.h
@@ -271,7 +271,7 @@ int lscpu_read_numas(struct lscpu_cxt *cxt);
 void lscpu_free_caches(struct lscpu_cache *caches, size_t n);
 void lscpu_sort_caches(struct lscpu_cache *caches, size_t n);
 
-size_t lscpu_get_cache_full_size(struct lscpu_cxt *cxt, const char *name);
+size_t lscpu_get_cache_full_size(struct lscpu_cxt *cxt, const char *name, int *instances);
 struct lscpu_cache *lscpu_cpu_get_cache(struct lscpu_cxt *cxt,
                                 struct lscpu_cpu *cpu, const char *name);